diff --git a/.gitignore b/.gitignore index 82146fa..6f0d9df 100644 --- a/.gitignore +++ b/.gitignore @@ -2,17 +2,19 @@ .vscode -ucloud/docs/build +uncloud/docs/build logs.txt -ucloud.egg-info +uncloud.egg-info # run artefacts default.etcd __pycache__ # build artefacts -ucloud/version.py +uncloud/version.py build/ venv/ dist/ + +*.iso diff --git a/bin/gen-version b/bin/gen-version new file mode 100755 index 0000000..06c3e22 --- /dev/null +++ b/bin/gen-version @@ -0,0 +1,29 @@ +#!/bin/sh +# -*- coding: utf-8 -*- +# +# 2019-2020 Nico Schottelius (nico-uncloud at schottelius.org) +# +# This file is part of uncloud. +# +# uncloud is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# uncloud is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with uncloud. If not, see . +# +# + + +# Wrapper for real script to allow execution from checkout +dir=${0%/*} + +# Ensure version is present - the bundled/shipped version contains a static version, +# the git version contains a dynamic version +printf "VERSION = \"%s\"\n" "$(git describe --tags --abbrev=0)" > ${dir}/../uncloud/version.py diff --git a/bin/ucloud b/bin/uncloud similarity index 90% rename from bin/ucloud rename to bin/uncloud index e178413..1c572d5 100755 --- a/bin/ucloud +++ b/bin/uncloud @@ -25,9 +25,9 @@ dir=${0%/*} # Ensure version is present - the bundled/shipped version contains a static version, # the git version contains a dynamic version -printf "VERSION = \"%s\"\n" "$(git describe)" > ${dir}/../ucloud/version.py +${dir}/gen-version libdir=$(cd "${dir}/../" && pwd -P) export PYTHONPATH="${libdir}" -"$dir/../scripts/ucloud" "$@" +"$dir/../scripts/uncloud" "$@" diff --git a/bin/uncloud-run-reinstall b/bin/uncloud-run-reinstall new file mode 100755 index 0000000..b211613 --- /dev/null +++ b/bin/uncloud-run-reinstall @@ -0,0 +1,29 @@ +#!/bin/sh +# -*- coding: utf-8 -*- +# +# 2012-2019 Nico Schottelius (nico-ucloud at schottelius.org) +# +# This file is part of ucloud. +# +# ucloud is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# ucloud is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with ucloud. If not, see . +# +# + +# Wrapper for real script to allow execution from checkout +dir=${0%/*} + +${dir}/gen-version; +pip uninstall -y uncloud >/dev/null +python setup.py install >/dev/null +${dir}/uncloud "$@" diff --git a/conf/uncloud.conf b/conf/uncloud.conf new file mode 100644 index 0000000..6a1b500 --- /dev/null +++ b/conf/uncloud.conf @@ -0,0 +1,13 @@ +[etcd] +url = localhost +port = 2379 +base_prefix = / +ca_cert +cert_cert +cert_key + +[client] +name = replace_me +realm = replace_me +seed = replace_me +api_server = http://localhost:5000 \ No newline at end of file diff --git a/ucloud/docs/Makefile b/docs/Makefile similarity index 93% rename from ucloud/docs/Makefile rename to docs/Makefile index 5e7ea85..246b56c 100644 --- a/ucloud/docs/Makefile +++ b/docs/Makefile @@ -7,7 +7,7 @@ SPHINXOPTS ?= SPHINXBUILD ?= sphinx-build SOURCEDIR = source/ BUILDDIR = build/ -DESTINATION=root@staticweb.ungleich.ch:/home/services/www/ungleichstatic/staticcms.ungleich.ch/www/ucloud/ +DESTINATION=root@staticweb.ungleich.ch:/home/services/www/ungleichstatic/staticcms.ungleich.ch/www/uncloud/ .PHONY: all build clean diff --git a/docs/README.md b/docs/README.md new file mode 100644 index 0000000..a5afbaa --- /dev/null +++ b/docs/README.md @@ -0,0 +1,12 @@ +# uncloud docs + +## Requirements +1. Python3 +2. Sphinx + +## Usage +Run `make build` to build docs. + +Run `make clean` to remove build directory. + +Run `make publish` to push build dir to https://ungleich.ch/ucloud/ \ No newline at end of file diff --git a/ucloud/__init__.py b/docs/__init__.py similarity index 100% rename from ucloud/__init__.py rename to docs/__init__.py diff --git a/ucloud/docs/__init__.py b/docs/source/__init__.py similarity index 100% rename from ucloud/docs/__init__.py rename to docs/source/__init__.py diff --git a/ucloud/docs/source/admin-guide b/docs/source/admin-guide.rst similarity index 72% rename from ucloud/docs/source/admin-guide rename to docs/source/admin-guide.rst index ec6597d..b62808d 100644 --- a/ucloud/docs/source/admin-guide +++ b/docs/source/admin-guide.rst @@ -56,40 +56,13 @@ To start host we created earlier, execute the following command ucloud host ungleich.ch -Create OS Image ---------------- +File & image scanners +-------------------------- -Create ucloud-init ready OS image (Optional) -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -This step is optional if you just want to test ucloud. However, sooner or later -you want to create OS images with ucloud-init to properly -contexualize VMs. - -1. Start a VM with OS image on which you want to install ucloud-init -2. Execute the following command on the started VM - - .. code-block:: sh - - apk add git - git clone https://code.ungleich.ch/ucloud/ucloud-init.git - cd ucloud-init - sh ./install.sh -3. Congratulations. Your image is now ucloud-init ready. - - -Upload Sample OS Image -~~~~~~~~~~~~~~~~~~~~~~ -Execute the following to get the sample OS image file. - -.. code-block:: sh - - mkdir /var/www/admin - (cd /var/www/admin && wget https://cloud.ungleich.ch/s/qTb5dFYW5ii8KsD/download) - -Run File Scanner and Image Scanner -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Currently, our uploaded file *alpine-untouched.qcow2* is not tracked by ucloud. We can only make -images from tracked files. So, we need to track the file by running File Scanner +Let's assume we have uploaded an *alpine-uploaded.qcow2* disk images to our +uncloud server. Currently, our *alpine-untouched.qcow2* is not tracked by +ucloud. We can only make images from tracked files. So, we need to track the +file by running File Scanner .. code-block:: sh diff --git a/ucloud/docs/source/conf.py b/docs/source/conf.py similarity index 90% rename from ucloud/docs/source/conf.py rename to docs/source/conf.py index 9b133f9..c8138a7 100644 --- a/ucloud/docs/source/conf.py +++ b/docs/source/conf.py @@ -17,9 +17,9 @@ # -- Project information ----------------------------------------------------- -project = 'ucloud' -copyright = '2019, ungleich' -author = 'ungleich' +project = "uncloud" +copyright = "2019, ungleich" +author = "ungleich" # -- General configuration --------------------------------------------------- @@ -27,12 +27,12 @@ author = 'ungleich' # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom # ones. extensions = [ - 'sphinx.ext.autodoc', - 'sphinx_rtd_theme', + "sphinx.ext.autodoc", + "sphinx_rtd_theme", ] # Add any paths that contain templates here, relative to this directory. -templates_path = ['_templates'] +templates_path = ["_templates"] # List of patterns, relative to source directory, that match files and # directories to ignore when looking for source files. @@ -50,4 +50,4 @@ html_theme = "sphinx_rtd_theme" # Add any paths that contain custom static files (such as style sheets) here, # relative to this directory. They are copied after the builtin static files, # so a file named "default.css" will overwrite the builtin "default.css". -html_static_path = ['_static'] +html_static_path = ["_static"] diff --git a/ucloud/docs/source/diagram-code/ucloud b/docs/source/diagram-code/ucloud similarity index 100% rename from ucloud/docs/source/diagram-code/ucloud rename to docs/source/diagram-code/ucloud diff --git a/docs/source/hacking.rst b/docs/source/hacking.rst new file mode 100644 index 0000000..1c750d6 --- /dev/null +++ b/docs/source/hacking.rst @@ -0,0 +1,36 @@ +Hacking +======= +Using uncloud in hacking (aka development) mode. + + +Get the code +------------ +.. code-block:: sh + :linenos: + + git clone https://code.ungleich.ch/uncloud/uncloud.git + + + +Install python requirements +--------------------------- +You need to have python3 installed. + +.. code-block:: sh + :linenos: + + cd uncloud! + python -m venv venv + . ./venv/bin/activate + ./bin/uncloud-run-reinstall + + + +Install os requirements +----------------------- +Install the following software packages: **dnsmasq**. + +If you already have a working IPv6 SLAAC and DNS setup, +this step can be skipped. + +Note that you need at least one /64 IPv6 network to run uncloud. diff --git a/ucloud/docs/source/images/ucloud.svg b/docs/source/images/ucloud.svg similarity index 100% rename from ucloud/docs/source/images/ucloud.svg rename to docs/source/images/ucloud.svg diff --git a/ucloud/docs/source/index.rst b/docs/source/index.rst similarity index 90% rename from ucloud/docs/source/index.rst rename to docs/source/index.rst index 879ac32..fad1f88 100644 --- a/ucloud/docs/source/index.rst +++ b/docs/source/index.rst @@ -11,12 +11,12 @@ Welcome to ucloud's documentation! :caption: Contents: introduction - user-guide setup-install + vm-images + user-guide admin-guide - user-guide/how-to-create-an-os-image-for-ucloud troubleshooting - + hacking Indices and tables ================== diff --git a/ucloud/docs/source/introduction.rst b/docs/source/introduction.rst similarity index 100% rename from ucloud/docs/source/introduction.rst rename to docs/source/introduction.rst diff --git a/ucloud/docs/source/misc/todo.rst b/docs/source/misc/todo.rst similarity index 100% rename from ucloud/docs/source/misc/todo.rst rename to docs/source/misc/todo.rst diff --git a/ucloud/docs/source/setup-install.rst b/docs/source/setup-install.rst similarity index 100% rename from ucloud/docs/source/setup-install.rst rename to docs/source/setup-install.rst diff --git a/ucloud/docs/source/theory/summary.rst b/docs/source/theory/summary.rst similarity index 100% rename from ucloud/docs/source/theory/summary.rst rename to docs/source/theory/summary.rst diff --git a/ucloud/docs/source/troubleshooting.rst b/docs/source/troubleshooting.rst similarity index 100% rename from ucloud/docs/source/troubleshooting.rst rename to docs/source/troubleshooting.rst diff --git a/ucloud/docs/source/user-guide.rst b/docs/source/user-guide.rst similarity index 100% rename from ucloud/docs/source/user-guide.rst rename to docs/source/user-guide.rst diff --git a/ucloud/docs/source/user-guide/how-to-create-an-os-image-for-ucloud.rst b/docs/source/user-guide/how-to-create-an-os-image-for-ucloud.rst similarity index 100% rename from ucloud/docs/source/user-guide/how-to-create-an-os-image-for-ucloud.rst rename to docs/source/user-guide/how-to-create-an-os-image-for-ucloud.rst diff --git a/docs/source/vm-images.rst b/docs/source/vm-images.rst new file mode 100644 index 0000000..4b2758a --- /dev/null +++ b/docs/source/vm-images.rst @@ -0,0 +1,66 @@ +VM images +================================== + +Overview +--------- + +ucloud tries to be least invasise towards VMs and only require +strictly necessary changes for running in a virtualised +environment. This includes configurations for: + +* Configuring the network +* Managing access via ssh keys +* Resizing the attached disk(s) + +Upstream images +--------------- + +The 'official' uncloud images are defined in the `uncloud/images +`_ repository. + +How to make you own Uncloud images +---------------------------------- + +.. note:: + It is fairly easy to create your own images for uncloud, as the common + operations (which are detailed below) can be automatically handled by the + `uncloud/uncloud-init `_ tool. + +Network configuration +~~~~~~~~~~~~~~~~~~~~~ +All VMs in ucloud are required to support IPv6. The primary network +configuration is always done using SLAAC. A VM thus needs only to be +configured to + +* accept router advertisements on all network interfaces +* use the router advertisements to configure the network interfaces +* accept the DNS entries from the router advertisements + + +Configuring SSH keys +~~~~~~~~~~~~~~~~~~~~ + +To be able to access the VM, ucloud support provisioning SSH keys. + +To accept ssh keys in your VM, request the URL +*http://metadata/ssh_keys*. Add the content to the appropriate user's +**authorized_keys** file. Below you find sample code to accomplish +this task: + +.. code-block:: sh + + tmp=$(mktemp) + curl -s http://metadata/ssk_keys > "$tmp" + touch ~/.ssh/authorized_keys # ensure it exists + cat ~/.ssh/authorized_keys >> "$tmp" + sort "$tmp" | uniq > ~/.ssh/authorized_keys + + +Disk resize +~~~~~~~~~~~ +In virtualised environments, the disk sizes might grow. The operating +system should detect disks that are bigger than the existing partition +table and resize accordingly. This task is os specific. + +ucloud does not support shrinking disks due to the complexity and +intra OS dependencies. diff --git a/scripts/ucloud b/scripts/ucloud deleted file mode 100755 index 7be6b24..0000000 --- a/scripts/ucloud +++ /dev/null @@ -1,59 +0,0 @@ -#!/usr/bin/env python3 - -import argparse -import multiprocessing as mp -import logging - -from os.path import join as join_path -from ucloud.sanity_checks import check - -if __name__ == "__main__": - arg_parser = argparse.ArgumentParser(prog='ucloud', - description='Open Source Cloud Management Software') - arg_parser.add_argument('component', - choices=['api', 'scheduler', 'host', - 'filescanner', 'imagescanner', - 'metadata']) - arg_parser.add_argument('component_args', nargs='*') - args = arg_parser.parse_args() - - logging.basicConfig( - level=logging.DEBUG, - filename=join_path("/", "etc", "ucloud", "log.txt"), - filemode="a", - format="%(name)s %(asctime)s: %(levelname)s - %(message)s", - datefmt="%d-%b-%y %H:%M:%S", - ) - try: - check() - - if args.component == 'api': - from ucloud.api.main import main - - main() - elif args.component == 'host': - from ucloud.host.main import main - - hostname = args.component_args - mp.set_start_method('spawn') - main(*hostname) - elif args.component == 'scheduler': - from ucloud.scheduler.main import main - - main() - elif args.component == 'filescanner': - from ucloud.filescanner.main import main - - main() - elif args.component == 'imagescanner': - from ucloud.imagescanner.main import main - - main() - elif args.component == 'metadata': - from ucloud.metadata.main import main - - main() - - except Exception as e: - logging.exception(e) - print(e) \ No newline at end of file diff --git a/scripts/uncloud b/scripts/uncloud new file mode 100755 index 0000000..7d38e42 --- /dev/null +++ b/scripts/uncloud @@ -0,0 +1,88 @@ +#!/usr/bin/env python3 +import logging +import sys +import importlib +import argparse +import os + +from etcd3.exceptions import ConnectionFailedError + +from uncloud.common import settings +from uncloud import UncloudException +from uncloud.common.cli import resolve_otp_credentials + +# Components that use etcd +ETCD_COMPONENTS = ['api', 'scheduler', 'host', 'filescanner', + 'imagescanner', 'metadata', 'configure', 'hack'] + +ALL_COMPONENTS = ETCD_COMPONENTS.copy() +ALL_COMPONENTS.append('oneshot') +#ALL_COMPONENTS.append('cli') + + +if __name__ == '__main__': + arg_parser = argparse.ArgumentParser() + subparsers = arg_parser.add_subparsers(dest='command') + + parent_parser = argparse.ArgumentParser(add_help=False) + parent_parser.add_argument('--debug', '-d', action='store_true', default=False, + help='More verbose logging') + parent_parser.add_argument('--conf-dir', '-c', help='Configuration directory', + default=os.path.expanduser('~/uncloud')) + + etcd_parser = argparse.ArgumentParser(add_help=False) + etcd_parser.add_argument('--etcd-host') + etcd_parser.add_argument('--etcd-port') + etcd_parser.add_argument('--etcd-ca-cert', help='CA that signed the etcd certificate') + etcd_parser.add_argument('--etcd-cert-cert', help='Path to client certificate') + etcd_parser.add_argument('--etcd-cert-key', help='Path to client certificate key') + + for component in ALL_COMPONENTS: + mod = importlib.import_module('uncloud.{}.main'.format(component)) + parser = getattr(mod, 'arg_parser') + + if component in ETCD_COMPONENTS: + subparsers.add_parser(name=parser.prog, parents=[parser, parent_parser, etcd_parser]) + else: + subparsers.add_parser(name=parser.prog, parents=[parser, parent_parser]) + + arguments = vars(arg_parser.parse_args()) + etcd_arguments = [key for key, value in arguments.items() if key.startswith('etcd_') and value] + etcd_arguments = { + 'etcd': { + key.replace('etcd_', ''): arguments[key] + for key in etcd_arguments + } + } + if not arguments['command']: + arg_parser.print_help() + else: + # Initializing Settings and resolving otp_credentials + # It is neccessary to resolve_otp_credentials after argument parsing is done because + # previously we were reading config file which was fixed to ~/uncloud/uncloud.conf and + # providing the default values for --name, --realm and --seed arguments from the values + # we read from file. But, now we are asking user about where the config file lives. So, + # to providing default value is not possible before parsing arguments. So, we are doing + # it after.. +# settings.settings = settings.Settings(arguments['conf_dir'], seed_value=etcd_arguments) +# resolve_otp_credentials(arguments) + + name = arguments.pop('command') + mod = importlib.import_module('uncloud.{}.main'.format(name)) + main = getattr(mod, 'main') + + if arguments['debug']: + logging.basicConfig(level=logging.DEBUG) + else: + logging.basicConfig(level=logging.INFO) + + log = logging.getLogger() + + try: + main(arguments) + except UncloudException as err: + log.error(err) +# except ConnectionFailedError as err: +# log.error('Cannot connect to etcd: {}'.format(err)) + except Exception as err: + log.exception(err) diff --git a/setup.py b/setup.py index 9a35f27..12da6b8 100644 --- a/setup.py +++ b/setup.py @@ -1,35 +1,50 @@ +import os + from setuptools import setup, find_packages with open("README.md", "r") as fh: long_description = fh.read() -setup(name='ucloud', - version='0.0.1', - description='All ucloud server components.', - url='https://code.ungleich.ch/ucloud/ucloud', - long_description=long_description, - long_description_content_type='text/markdown', - classifiers=[ - 'Development Status :: 3 - Alpha', - 'License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)', - 'Programming Language :: Python :: 3' - ], - author='ungleich', - author_email='technik@ungleich.ch', - packages=find_packages(), - install_requires=[ - 'requests', - 'python-decouple', - 'flask', - 'flask-restful', - 'bitmath', - 'pyotp', - 'sshtunnel', - 'sphinx', - 'pynetbox', - 'sphinx-rtd-theme', - 'etcd3_wrapper @ https://code.ungleich.ch/ungleich-public/etcd3_wrapper/repository/master/archive.tar.gz#egg=etcd3_wrapper', - 'etcd3 @ https://github.com/kragniz/python-etcd3/tarball/master#egg=etcd3', - ], - scripts=['scripts/ucloud'], - zip_safe=False) +try: + import uncloud.version + + version = uncloud.version.VERSION +except: + import subprocess + + c = subprocess.check_output(["git", "describe"]) + version = c.decode("utf-8").strip() + + +setup( + name="uncloud", + version=version, + description="uncloud cloud management", + url="https://code.ungleich.ch/uncloud/uncloud", + long_description=long_description, + long_description_content_type="text/markdown", + classifiers=[ + "Development Status :: 3 - Alpha", + "License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)", + "Programming Language :: Python :: 3", + ], + author="ungleich", + author_email="technik@ungleich.ch", + packages=find_packages(), + install_requires=[ + "requests", + "Flask>=1.1.1", + "flask-restful", + "bitmath", + "pyotp", + "pynetbox", + "colorama", + "etcd3 @ https://github.com/kragniz/python-etcd3/tarball/master#egg=etcd3", + "marshmallow" + ], + scripts=["scripts/uncloud"], + data_files=[ + (os.path.expanduser("~/uncloud/"), ["conf/uncloud.conf"]) + ], + zip_safe=False, +) diff --git a/ucloud/api/create_image_store.py b/ucloud/api/create_image_store.py deleted file mode 100755 index 17fa63c..0000000 --- a/ucloud/api/create_image_store.py +++ /dev/null @@ -1,16 +0,0 @@ -import json -import os - -from uuid import uuid4 - -from ucloud.config import etcd_client, env_vars - -data = { - "is_public": True, - "type": "ceph", - "name": "images", - "description": "first ever public image-store", - "attributes": {"list": [], "key": [], "pool": "images"}, -} - -etcd_client.put(os.path.join(env_vars.get('IMAGE_STORE_PREFIX'), uuid4().hex), json.dumps(data)) diff --git a/ucloud/api/main.py b/ucloud/api/main.py deleted file mode 100644 index 1475fb0..0000000 --- a/ucloud/api/main.py +++ /dev/null @@ -1,517 +0,0 @@ -import json -import pynetbox - -from uuid import uuid4 -from os.path import join as join_path - -from flask import Flask, request -from flask_restful import Resource, Api - -from ucloud.common import counters -from ucloud.common.vm import VMStatus -from ucloud.common.request import RequestEntry, RequestType -from ucloud.config import (etcd_client, request_pool, vm_pool, host_pool, env_vars, image_storage_handler) -from . import schemas -from .helper import generate_mac, mac2ipv6 -from . import logger - -app = Flask(__name__) -api = Api(app) - - -class CreateVM(Resource): - """API Request to Handle Creation of VM""" - - @staticmethod - def post(): - data = request.json - validator = schemas.CreateVMSchema(data) - if validator.is_valid(): - vm_uuid = uuid4().hex - vm_key = join_path(env_vars.get("VM_PREFIX"), vm_uuid) - specs = { - "cpu": validator.specs["cpu"], - "ram": validator.specs["ram"], - "os-ssd": validator.specs["os-ssd"], - "hdd": validator.specs["hdd"], - } - macs = [generate_mac() for _ in range(len(data["network"]))] - tap_ids = [counters.increment_etcd_counter(etcd_client, "/v1/counter/tap") - for _ in range(len(data["network"]))] - vm_entry = { - "name": data["vm_name"], - "owner": data["name"], - "owner_realm": data["realm"], - "specs": specs, - "hostname": "", - "status": VMStatus.stopped, - "image_uuid": validator.image_uuid, - "log": [], - "vnc_socket": "", - "network": list(zip(data["network"], macs, tap_ids)), - "metadata": {"ssh-keys": []}, - } - etcd_client.put(vm_key, vm_entry, value_in_json=True) - - # Create ScheduleVM Request - r = RequestEntry.from_scratch( - type=RequestType.ScheduleVM, uuid=vm_uuid, - request_prefix=env_vars.get("REQUEST_PREFIX") - ) - request_pool.put(r) - - return {"message": "VM Creation Queued"}, 200 - return validator.get_errors(), 400 - - -class VmStatus(Resource): - @staticmethod - def get(): - data = request.json - validator = schemas.VMStatusSchema(data) - if validator.is_valid(): - vm = vm_pool.get( - join_path(env_vars.get("VM_PREFIX"), data["uuid"]) - ) - vm_value = vm.value.copy() - vm_value["ip"] = [] - for network_mac_and_tap in vm.network: - network_name, mac, tap = network_mac_and_tap - network = etcd_client.get( - join_path( - env_vars.get("NETWORK_PREFIX"), - data["name"], - network_name, - ), - value_in_json=True, - ) - ipv6_addr = network.value.get("ipv6").split("::")[0] + "::" - vm_value["ip"].append(mac2ipv6(mac, ipv6_addr)) - vm.value = vm_value - return vm.value - else: - return validator.get_errors(), 400 - - -class CreateImage(Resource): - @staticmethod - def post(): - data = request.json - validator = schemas.CreateImageSchema(data) - if validator.is_valid(): - file_entry = etcd_client.get( - join_path(env_vars.get("FILE_PREFIX"), data["uuid"]) - ) - file_entry_value = json.loads(file_entry.value) - - image_entry_json = { - "status": "TO_BE_CREATED", - "owner": file_entry_value["owner"], - "filename": file_entry_value["filename"], - "name": data["name"], - "store_name": data["image_store"], - "visibility": "public", - } - etcd_client.put( - join_path(env_vars.get("IMAGE_PREFIX"), data["uuid"]), - json.dumps(image_entry_json), - ) - - return {"message": "Image queued for creation."} - return validator.get_errors(), 400 - - -class ListPublicImages(Resource): - @staticmethod - def get(): - images = etcd_client.get_prefix( - env_vars.get("IMAGE_PREFIX"), value_in_json=True - ) - r = { - "images": [] - } - for image in images: - image_key = "{}:{}".format( - image.value["store_name"], image.value["name"] - ) - r["images"].append( - {"name": image_key, "status": image.value["status"]} - ) - return r, 200 - - -class VMAction(Resource): - @staticmethod - def post(): - data = request.json - validator = schemas.VmActionSchema(data) - - if validator.is_valid(): - vm_entry = vm_pool.get( - join_path(env_vars.get("VM_PREFIX"), data["uuid"]) - ) - action = data["action"] - - if action == "start": - action = "schedule" - - if action == "delete" and vm_entry.hostname == "": - if image_storage_handler.is_vm_image_exists(vm_entry.uuid): - r_status = image_storage_handler.delete_vm_image(vm_entry.uuid) - if r_status: - etcd_client.client.delete(vm_entry.key) - return {"message": "VM successfully deleted"} - else: - logger.error("Some Error Occurred while deleting VM") - return {"message": "VM deletion unsuccessfull"} - else: - etcd_client.client.delete(vm_entry.key) - return {"message": "VM successfully deleted"} - - r = RequestEntry.from_scratch( - type="{}VM".format(action.title()), - uuid=data["uuid"], - hostname=vm_entry.hostname, - request_prefix=env_vars.get("REQUEST_PREFIX") - ) - request_pool.put(r) - return {"message": "VM {} Queued".format(action.title())}, 200 - else: - return validator.get_errors(), 400 - - -class VMMigration(Resource): - @staticmethod - def post(): - data = request.json - validator = schemas.VmMigrationSchema(data) - - if validator.is_valid(): - vm = vm_pool.get(data["uuid"]) - - r = RequestEntry.from_scratch( - type=RequestType.ScheduleVM, - uuid=vm.uuid, - destination=join_path( - env_vars.get("HOST_PREFIX"), validator.destination.value - ), - migration=True, - request_prefix=env_vars.get("REQUEST_PREFIX") - ) - request_pool.put(r) - return {"message": "VM Migration Initialization Queued"}, 200 - else: - return validator.get_errors(), 400 - - -class ListUserVM(Resource): - @staticmethod - def get(): - data = request.json - validator = schemas.OTPSchema(data) - - if validator.is_valid(): - vms = etcd_client.get_prefix( - env_vars.get("VM_PREFIX"), value_in_json=True - ) - return_vms = [] - user_vms = filter(lambda v: v.value["owner"] == data["name"], vms) - for vm in user_vms: - return_vms.append( - { - "name": vm.value["name"], - "vm_uuid": vm.key.split("/")[-1], - "specs": vm.value["specs"], - "status": vm.value["status"], - "hostname": vm.value["hostname"], - # "mac": vm.value["mac"], - "vnc_socket": None - if vm.value.get("vnc_socket", None) is None - else vm.value["vnc_socket"], - } - ) - if return_vms: - return {"message": return_vms}, 200 - return {"message": "No VM found"}, 404 - - else: - return validator.get_errors(), 400 - - -class ListUserFiles(Resource): - @staticmethod - def get(): - data = request.json - validator = schemas.OTPSchema(data) - - if validator.is_valid(): - files = etcd_client.get_prefix( - env_vars.get("FILE_PREFIX"), value_in_json=True - ) - return_files = [] - user_files = list( - filter(lambda f: f.value["owner"] == data["name"], files) - ) - for file in user_files: - return_files.append( - { - "filename": file.value["filename"], - "uuid": file.key.split("/")[-1], - } - ) - return {"message": return_files}, 200 - else: - return validator.get_errors(), 400 - - -class CreateHost(Resource): - @staticmethod - def post(): - data = request.json - validator = schemas.CreateHostSchema(data) - if validator.is_valid(): - host_key = join_path(env_vars.get("HOST_PREFIX"), uuid4().hex) - host_entry = { - "specs": data["specs"], - "hostname": data["hostname"], - "status": "DEAD", - "last_heartbeat": "", - } - etcd_client.put(host_key, host_entry, value_in_json=True) - - return {"message": "Host Created"}, 200 - - return validator.get_errors(), 400 - - -class ListHost(Resource): - @staticmethod - def get(): - hosts = host_pool.hosts - r = { - host.key: { - "status": host.status, - "specs": host.specs, - "hostname": host.hostname, - } - for host in hosts - } - return r, 200 - - -class GetSSHKeys(Resource): - @staticmethod - def get(): - data = request.json - validator = schemas.GetSSHSchema(data) - if validator.is_valid(): - if not validator.key_name.value: - - # {user_prefix}/{realm}/{name}/key/ - etcd_key = join_path( - env_vars.get('USER_PREFIX'), - data["realm"], - data["name"], - "key", - ) - etcd_entry = etcd_client.get_prefix( - etcd_key, value_in_json=True - ) - - keys = { - key.key.split("/")[-1]: key.value for key in etcd_entry - } - return {"keys": keys} - else: - - # {user_prefix}/{realm}/{name}/key/{key_name} - etcd_key = join_path( - env_vars.get('USER_PREFIX'), - data["realm"], - data["name"], - "key", - data["key_name"], - ) - etcd_entry = etcd_client.get(etcd_key, value_in_json=True) - - if etcd_entry: - return { - "keys": { - etcd_entry.key.split("/")[-1]: etcd_entry.value - } - } - else: - return {"keys": {}} - else: - return validator.get_errors(), 400 - - -class AddSSHKey(Resource): - @staticmethod - def post(): - data = request.json - validator = schemas.AddSSHSchema(data) - if validator.is_valid(): - - # {user_prefix}/{realm}/{name}/key/{key_name} - etcd_key = join_path( - env_vars.get("USER_PREFIX"), - data["realm"], - data["name"], - "key", - data["key_name"], - ) - etcd_entry = etcd_client.get(etcd_key, value_in_json=True) - if etcd_entry: - return { - "message": "Key with name '{}' already exists".format( - data["key_name"] - ) - } - else: - # Key Not Found. It implies user' haven't added any key yet. - etcd_client.put(etcd_key, data["key"], value_in_json=True) - return {"message": "Key added successfully"} - else: - return validator.get_errors(), 400 - - -class RemoveSSHKey(Resource): - @staticmethod - def get(): - data = request.json - validator = schemas.RemoveSSHSchema(data) - if validator.is_valid(): - - # {user_prefix}/{realm}/{name}/key/{key_name} - etcd_key = join_path( - env_vars.get("USER_PREFIX"), - data["realm"], - data["name"], - "key", - data["key_name"], - ) - etcd_entry = etcd_client.get(etcd_key, value_in_json=True) - if etcd_entry: - etcd_client.client.delete(etcd_key) - return {"message": "Key successfully removed."} - else: - return { - "message": "No Key with name '{}' Exists at all.".format( - data["key_name"] - ) - } - else: - return validator.get_errors(), 400 - - -class CreateNetwork(Resource): - @staticmethod - def post(): - data = request.json - validator = schemas.CreateNetwork(data) - - if validator.is_valid(): - - network_entry = { - "id": counters.increment_etcd_counter( - etcd_client, "/v1/counter/vxlan" - ), - "type": data["type"], - } - if validator.user.value: - nb = pynetbox.api( - url=env_vars.get("NETBOX_URL"), - token=env_vars.get("NETBOX_TOKEN"), - ) - nb_prefix = nb.ipam.prefixes.get( - prefix=env_vars.get("PREFIX") - ) - - prefix = nb_prefix.available_prefixes.create( - data={ - "prefix_length": env_vars.get( - "PREFIX_LENGTH", cast=int - ), - "description": '{}\'s network "{}"'.format( - data["name"], data["network_name"] - ), - "is_pool": True, - } - ) - network_entry["ipv6"] = prefix["prefix"] - else: - network_entry["ipv6"] = "fd00::/64" - - network_key = join_path( - env_vars.get("NETWORK_PREFIX"), - data["name"], - data["network_name"], - ) - etcd_client.put(network_key, network_entry, value_in_json=True) - return {"message": "Network successfully added."} - else: - return validator.get_errors(), 400 - - -class ListUserNetwork(Resource): - @staticmethod - def get(): - data = request.json - validator = schemas.OTPSchema(data) - - if validator.is_valid(): - prefix = join_path( - env_vars.get("NETWORK_PREFIX"), data["name"] - ) - networks = etcd_client.get_prefix(prefix, value_in_json=True) - user_networks = [] - for net in networks: - net.value["name"] = net.key.split("/")[-1] - user_networks.append(net.value) - return {"networks": user_networks}, 200 - else: - return validator.get_errors(), 400 - - -api.add_resource(CreateVM, "/vm/create") -api.add_resource(VmStatus, "/vm/status") - -api.add_resource(VMAction, "/vm/action") -api.add_resource(VMMigration, "/vm/migrate") - -api.add_resource(CreateImage, "/image/create") -api.add_resource(ListPublicImages, "/image/list-public") - -api.add_resource(ListUserVM, "/user/vms") -api.add_resource(ListUserFiles, "/user/files") -api.add_resource(ListUserNetwork, "/user/networks") - -api.add_resource(AddSSHKey, "/user/add-ssh") -api.add_resource(RemoveSSHKey, "/user/remove-ssh") -api.add_resource(GetSSHKeys, "/user/get-ssh") - -api.add_resource(CreateHost, "/host/create") -api.add_resource(ListHost, "/host/list") - -api.add_resource(CreateNetwork, "/network/create") - - -def main(): - image_stores = list(etcd_client.get_prefix(env_vars.get('IMAGE_STORE_PREFIX'), value_in_json=True)) - if len(image_stores) == 0: - data = { - "is_public": True, - "type": "ceph", - "name": "images", - "description": "first ever public image-store", - "attributes": {"list": [], "key": [], "pool": "images"}, - } - - etcd_client.put(join_path(env_vars.get('IMAGE_STORE_PREFIX'), uuid4().hex), json.dumps(data)) - - app.run(host="::", debug=True) - - -if __name__ == "__main__": - main() diff --git a/ucloud/common/helpers.py b/ucloud/common/helpers.py deleted file mode 100644 index 1bdf0b4..0000000 --- a/ucloud/common/helpers.py +++ /dev/null @@ -1,54 +0,0 @@ -import logging -import socket -import requests -import json - -from ipaddress import ip_address - -from os.path import join as join_path - - -def create_package_loggers(packages, base_path, mode="a"): - loggers = {} - for pkg in packages: - logger = logging.getLogger(pkg) - logger_handler = logging.FileHandler( - join_path(base_path, "{}.txt".format(pkg)), - mode=mode - ) - logger.setLevel(logging.DEBUG) - logger_handler.setFormatter(logging.Formatter(fmt="%(asctime)s: %(levelname)s - %(message)s", - datefmt="%d-%b-%y %H:%M:%S")) - logger.addHandler(logger_handler) - loggers[pkg] = logger - - -# TODO: Should be removed as soon as migration -# mechanism is finalized inside ucloud -def get_ipv4_address(): - # If host is connected to internet - # Return IPv4 address of machine - # Otherwise, return 127.0.0.1 - with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as s: - try: - s.connect(("8.8.8.8", 80)) - except socket.timeout: - address = "127.0.0.1" - except Exception as e: - logging.getLogger().exception(e) - address = "127.0.0.1" - else: - address = s.getsockname()[0] - - return address - - -def get_ipv6_address(): - try: - r = requests.get("https://api6.ipify.org?format=json") - content = json.loads(r.content.decode("utf-8")) - ip = ip_address(content["ip"]).exploded - except Exception as e: - logging.exception(e) - else: - return ip diff --git a/ucloud/config.py b/ucloud/config.py deleted file mode 100644 index 7c141a3..0000000 --- a/ucloud/config.py +++ /dev/null @@ -1,41 +0,0 @@ -from etcd3_wrapper import Etcd3Wrapper - -from ucloud.common.host import HostPool -from ucloud.common.request import RequestPool -from ucloud.common.vm import VmPool -from ucloud.common.storage_handlers import FileSystemBasedImageStorageHandler, CEPHBasedImageStorageHandler -from decouple import Config, RepositoryEnv, RepositoryEmpty - - -# Try importing config, but don't fail if it does not exist -try: - env_vars = Config(RepositoryEnv('/etc/ucloud/ucloud.conf')) -except FileNotFoundError: - env_vars = Config(RepositoryEmpty()) - - -etcd_wrapper_args = () -etcd_wrapper_kwargs = { - 'host': env_vars.get('ETCD_URL', 'localhost'), - 'port': env_vars.get('ETCD_PORT', 2379), - 'ca_cert': env_vars.get('CA_CERT', None), - 'cert_cert': env_vars.get('CERT_CERT', None), - 'cert_key': env_vars.get('CERT_KEY', None) -} - -etcd_client = Etcd3Wrapper(*etcd_wrapper_args, **etcd_wrapper_kwargs) - -host_pool = HostPool(etcd_client, env_vars.get('HOST_PREFIX')) -vm_pool = VmPool(etcd_client, env_vars.get('VM_PREFIX')) -request_pool = RequestPool(etcd_client, env_vars.get('REQUEST_PREFIX')) - -running_vms = [] - -__storage_backend = env_vars.get("STORAGE_BACKEND") -if __storage_backend == "filesystem": - image_storage_handler = FileSystemBasedImageStorageHandler(vm_base=env_vars.get("VM_DIR"), - image_base=env_vars.get("IMAGE_DIR")) -elif __storage_backend == "ceph": - image_storage_handler = CEPHBasedImageStorageHandler(vm_base="ssd", image_base="ssd") -else: - raise Exception("Unknown Image Storage Handler") diff --git a/ucloud/filescanner/main.py b/ucloud/filescanner/main.py deleted file mode 100755 index b70cb5b..0000000 --- a/ucloud/filescanner/main.py +++ /dev/null @@ -1,126 +0,0 @@ -import glob -import os -import pathlib -import subprocess as sp -import time -from uuid import uuid4 - -from . import logger -from ucloud.config import env_vars, etcd_client - - -def getxattr(file, attr): - """Get specified user extended attribute (arg:attr) of a file (arg:file)""" - try: - attr = "user." + attr - value = sp.check_output(['getfattr', file, - '--name', attr, - '--only-values', - '--absolute-names'], stderr=sp.DEVNULL) - value = value.decode("utf-8") - except sp.CalledProcessError as e: - logger.exception(e) - value = None - - return value - - -def setxattr(file, attr, value): - """Set specified user extended attribute (arg:attr) equal to (arg:value) - of a file (arg:file)""" - - attr = "user." + attr - sp.check_output(['setfattr', file, - '--name', attr, - '--value', str(value)]) - - -def sha512sum(file: str): - """Use sha512sum utility to compute sha512 sum of arg:file - - IF arg:file does not exists: - raise FileNotFoundError exception - ELSE IF sum successfully computer: - return computed sha512 sum - ELSE: - return None - """ - if not isinstance(file, str): raise TypeError - try: - output = sp.check_output(["sha512sum", file], stderr=sp.PIPE) - except sp.CalledProcessError as e: - error = e.stderr.decode("utf-8") - if "No such file or directory" in error: - raise FileNotFoundError from None - else: - output = output.decode("utf-8").strip() - output = output.split(" ") - return output[0] - return None - - -try: - sp.check_output(['which', 'getfattr']) - sp.check_output(['which', 'setfattr']) -except Exception as e: - logger.exception(e) - print('Make sure you have getfattr and setfattr available') - exit(1) - - -def main(): - BASE_DIR = env_vars.get("BASE_DIR") - - FILE_PREFIX = env_vars.get("FILE_PREFIX") - - # Recursively Get All Files and Folder below BASE_DIR - files = glob.glob("{}/**".format(BASE_DIR), recursive=True) - - # Retain only Files - files = list(filter(os.path.isfile, files)) - - untracked_files = list( - filter(lambda f: not bool(getxattr(f, "user.utracked")), files) - ) - - tracked_files = list( - filter(lambda f: f not in untracked_files, files) - ) - for file in untracked_files: - file_id = uuid4() - - # Get Username - owner = pathlib.Path(file).parts[3] - # Get Creation Date of File - # Here, we are assuming that ctime is creation time - # which is mostly not true. - creation_date = time.ctime(os.stat(file).st_ctime) - - # Get File Size - size = os.path.getsize(file) - - # Compute sha512 sum - sha_sum = sha512sum(file) - - # File Path excluding base and username - file_path = pathlib.Path(file).parts[4:] - file_path = os.path.join(*file_path) - - # Create Entry - entry_key = os.path.join(FILE_PREFIX, str(file_id)) - entry_value = { - "filename": file_path, - "owner": owner, - "sha512sum": sha_sum, - "creation_date": creation_date, - "size": size - } - - print("Tracking {}".format(file)) - # Insert Entry - etcd_client.put(entry_key, entry_value, value_in_json=True) - setxattr(file, "user.utracked", True) - - -if __name__ == "__main__": - main() diff --git a/ucloud/host/helper.py b/ucloud/host/helper.py deleted file mode 100644 index edcb82d..0000000 --- a/ucloud/host/helper.py +++ /dev/null @@ -1,13 +0,0 @@ -import socket -from contextlib import closing - - -def find_free_port(): - with closing(socket.socket(socket.AF_INET, socket.SOCK_STREAM)) as s: - try: - s.bind(('', 0)) - s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) - except Exception: - return None - else: - return s.getsockname()[1] diff --git a/ucloud/host/main.py b/ucloud/host/main.py deleted file mode 100755 index ccf0a8d..0000000 --- a/ucloud/host/main.py +++ /dev/null @@ -1,143 +0,0 @@ -import argparse -import multiprocessing as mp -import time - -from etcd3_wrapper import Etcd3Wrapper - -from ucloud.common.request import RequestEntry, RequestType -from ucloud.config import (vm_pool, request_pool, - etcd_client, running_vms, - etcd_wrapper_args, etcd_wrapper_kwargs, - HostPool, env_vars) - -from .helper import find_free_port -from . import virtualmachine -from ucloud.host import logger - - -def update_heartbeat(hostname): - """Update Last HeartBeat Time for :param hostname: in etcd""" - client = Etcd3Wrapper(*etcd_wrapper_args, **etcd_wrapper_kwargs) - host_pool = HostPool(client, env_vars.get('HOST_PREFIX')) - this_host = next(filter(lambda h: h.hostname == hostname, host_pool.hosts), None) - - while True: - this_host.update_heartbeat() - host_pool.put(this_host) - time.sleep(10) - - -def maintenance(host): - # To capture vm running according to running_vms list - - # This is to capture successful migration of a VM. - # Suppose, this host is running "vm1" and user initiated - # request to migrate this "vm1" to some other host. On, - # successful migration the destination host would set - # the vm hostname to itself. Thus, we are checking - # whether this host vm is successfully migrated. If yes - # then we shutdown "vm1" on this host. - - to_be_removed = [] - for running_vm in running_vms: - with vm_pool.get_put(running_vm.key) as vm_entry: - if vm_entry.hostname != host.key and not vm_entry.in_migration: - running_vm.handle.shutdown() - logger.info("VM migration not completed successfully.") - to_be_removed.append(running_vm) - - for r in to_be_removed: - running_vms.remove(r) - - # To check vm running according to etcd entries - alleged_running_vms = vm_pool.by_status("RUNNING", vm_pool.by_host(host.key)) - - for vm_entry in alleged_running_vms: - _vm = virtualmachine.get_vm(running_vms, vm_entry.key) - # Whether, the allegedly running vm is in our - # running_vms list or not if it is said to be - # running on this host but it is not then we - # need to shut it down - - # This is to capture poweroff/shutdown of a VM - # initiated by user inside VM. OR crash of VM by some - # user running process - if (_vm and not _vm.handle.is_running()) or not _vm: - logger.debug("_vm = %s, is_running() = %s" % (_vm, _vm.handle.is_running())) - vm_entry.add_log("""{} is not running but is said to be running. - So, shutting it down and declare it killed""".format(vm_entry.key)) - vm_entry.declare_killed() - vm_pool.put(vm_entry) - if _vm: - running_vms.remove(_vm) - - -def main(hostname): - heartbeat_updating_process = mp.Process(target=update_heartbeat, args=(hostname,)) - - host_pool = HostPool(etcd_client, env_vars.get('HOST_PREFIX')) - host = next(filter(lambda h: h.hostname == hostname, host_pool.hosts), None) - assert host is not None, "No such host with name = {}".format(hostname) - - try: - heartbeat_updating_process.start() - except Exception as e: - logger.info("No Need To Go Further. Our heartbeat updating mechanism is not working") - logger.exception(e) - exit(-1) - - logger.info("%s Session Started %s", '*' * 5, '*' * 5) - - # It is seen that under heavy load, timeout event doesn't come - # in a predictive manner (which is intentional because we give - # higher priority to customer's requests) which delays heart - # beat update which in turn misunderstood by scheduler that the - # host is dead when it is actually alive. So, to ensure that we - # update the heart beat in a predictive manner we start Heart - # beat updating mechanism in separated thread - - for events_iterator in [ - etcd_client.get_prefix(env_vars.get('REQUEST_PREFIX'), value_in_json=True), - etcd_client.watch_prefix(env_vars.get('REQUEST_PREFIX'), timeout=10, value_in_json=True), - ]: - for request_event in events_iterator: - request_event = RequestEntry(request_event) - - if request_event.type == "TIMEOUT": - maintenance(host) - continue - - # If the event is directed toward me OR I am destination of a InitVMMigration - if request_event.hostname == host.key or request_event.destination == host.key: - logger.debug("VM Request: %s", request_event) - - request_pool.client.client.delete(request_event.key) - vm_entry = vm_pool.get(request_event.uuid) - - if vm_entry: - if request_event.type == RequestType.StartVM: - virtualmachine.start(vm_entry) - - elif request_event.type == RequestType.StopVM: - virtualmachine.stop(vm_entry) - - elif request_event.type == RequestType.DeleteVM: - virtualmachine.delete(vm_entry) - - elif request_event.type == RequestType.InitVMMigration: - virtualmachine.start(vm_entry, host.key, find_free_port()) - - elif request_event.type == RequestType.TransferVM: - virtualmachine.transfer(request_event) - else: - logger.info("VM Entry missing") - - logger.info("Running VMs %s", running_vms) - - -if __name__ == "__main__": - argparser = argparse.ArgumentParser() - argparser.add_argument("hostname", help="Name of this host. e.g /v1/host/1") - args = argparser.parse_args() - mp.set_start_method('spawn') - main(args.hostname) diff --git a/ucloud/host/qmp/__init__.py b/ucloud/host/qmp/__init__.py deleted file mode 100755 index 775b397..0000000 --- a/ucloud/host/qmp/__init__.py +++ /dev/null @@ -1,537 +0,0 @@ -# QEMU library -# -# Copyright (C) 2015-2016 Red Hat Inc. -# Copyright (C) 2012 IBM Corp. -# -# Authors: -# Fam Zheng -# -# This work is licensed under the terms of the GNU GPL, version 2. See -# the COPYING file in the top-level directory. -# -# Based on qmp.py. -# - -import errno -import logging -import os -import shutil -import socket -import subprocess -import tempfile - -from . import qmp - -LOG = logging.getLogger(__name__) - -# Mapping host architecture to any additional architectures it can -# support which often includes its 32 bit cousin. -ADDITIONAL_ARCHES = { - "x86_64": "i386", - "aarch64": "armhf" -} - - -def kvm_available(target_arch=None): - host_arch = os.uname()[4] - if target_arch and target_arch != host_arch: - if target_arch != ADDITIONAL_ARCHES.get(host_arch): - return False - return os.access("/dev/kvm", os.R_OK | os.W_OK) - - -class QEMUMachineError(Exception): - """ - Exception called when an error in QEMUMachine happens. - """ - - -class QEMUMachineAddDeviceError(QEMUMachineError): - """ - Exception raised when a request to add a device can not be fulfilled - - The failures are caused by limitations, lack of information or conflicting - requests on the QEMUMachine methods. This exception does not represent - failures reported by the QEMU binary itself. - """ - - -class MonitorResponseError(qmp.QMPError): - """ - Represents erroneous QMP monitor reply - """ - - def __init__(self, reply): - try: - desc = reply["error"]["desc"] - except KeyError: - desc = reply - super(MonitorResponseError, self).__init__(desc) - self.reply = reply - - -class QEMUMachine(object): - """ - A QEMU VM - - Use this object as a context manager to ensure the QEMU process terminates:: - - with VM(binary) as vm: - ... - # vm is guaranteed to be shut down here - """ - - def __init__(self, binary, args=None, wrapper=None, name=None, - test_dir="/var/tmp", monitor_address=None, - socket_scm_helper=None): - ''' - Initialize a QEMUMachine - - @param binary: path to the qemu binary - @param args: list of extra arguments - @param wrapper: list of arguments used as prefix to qemu binary - @param name: prefix for socket and log file names (default: qemu-PID) - @param test_dir: where to create socket and log file - @param monitor_address: address for QMP monitor - @param socket_scm_helper: helper program, required for send_fd_scm() - @note: Qemu process is not started until launch() is used. - ''' - if args is None: - args = [] - if wrapper is None: - wrapper = [] - if name is None: - name = "qemu-%d" % os.getpid() - self._name = name - self._monitor_address = monitor_address - self._vm_monitor = None - self._qemu_log_path = None - self._qemu_log_file = None - self._popen = None - self._binary = binary - self._args = list(args) # Force copy args in case we modify them - self._wrapper = wrapper - self._events = [] - self._iolog = None - self._socket_scm_helper = socket_scm_helper - self._qmp = None - self._qemu_full_args = None - self._test_dir = test_dir - self._temp_dir = None - self._launched = False - self._machine = None - self._console_set = False - self._console_device_type = None - self._console_address = None - self._console_socket = None - - # just in case logging wasn't configured by the main script: - logging.basicConfig(level=logging.DEBUG) - - def __enter__(self): - return self - - def __exit__(self, exc_type, exc_val, exc_tb): - self.shutdown() - return False - - # This can be used to add an unused monitor instance. - def add_monitor_null(self): - self._args.append('-monitor') - self._args.append('null') - - def add_fd(self, fd, fdset, opaque, opts=''): - """ - Pass a file descriptor to the VM - """ - options = ['fd=%d' % fd, - 'set=%d' % fdset, - 'opaque=%s' % opaque] - if opts: - options.append(opts) - - # This did not exist before 3.4, but since then it is - # mandatory for our purpose - if hasattr(os, 'set_inheritable'): - os.set_inheritable(fd, True) - - self._args.append('-add-fd') - self._args.append(','.join(options)) - return self - - # Exactly one of fd and file_path must be given. - # (If it is file_path, the helper will open that file and pass its - # own fd) - def send_fd_scm(self, fd=None, file_path=None): - # In iotest.py, the qmp should always use unix socket. - assert self._qmp.is_scm_available() - if self._socket_scm_helper is None: - raise QEMUMachineError("No path to socket_scm_helper set") - if not os.path.exists(self._socket_scm_helper): - raise QEMUMachineError("%s does not exist" % - self._socket_scm_helper) - - # This did not exist before 3.4, but since then it is - # mandatory for our purpose - if hasattr(os, 'set_inheritable'): - os.set_inheritable(self._qmp.get_sock_fd(), True) - if fd is not None: - os.set_inheritable(fd, True) - - fd_param = ["%s" % self._socket_scm_helper, - "%d" % self._qmp.get_sock_fd()] - - if file_path is not None: - assert fd is None - fd_param.append(file_path) - else: - assert fd is not None - fd_param.append(str(fd)) - - devnull = open(os.path.devnull, 'rb') - proc = subprocess.Popen(fd_param, stdin=devnull, stdout=subprocess.PIPE, - stderr=subprocess.STDOUT, close_fds=False) - output = proc.communicate()[0] - if output: - LOG.debug(output) - - return proc.returncode - - @staticmethod - def _remove_if_exists(path): - """ - Remove file object at path if it exists - """ - try: - os.remove(path) - except OSError as exception: - if exception.errno == errno.ENOENT: - return - raise - - def is_running(self): - return self._popen is not None and self._popen.poll() is None - - def exitcode(self): - if self._popen is None: - return None - return self._popen.poll() - - def get_pid(self): - if not self.is_running(): - return None - return self._popen.pid - - def _load_io_log(self): - if self._qemu_log_path is not None: - with open(self._qemu_log_path, "r") as iolog: - self._iolog = iolog.read() - - def _base_args(self): - if isinstance(self._monitor_address, tuple): - moncdev = "socket,id=mon,host=%s,port=%s" % ( - self._monitor_address[0], - self._monitor_address[1]) - else: - moncdev = 'socket,id=mon,path=%s' % self._vm_monitor - args = ['-chardev', moncdev, - '-mon', 'chardev=mon,mode=control'] - if self._machine is not None: - args.extend(['-machine', self._machine]) - if self._console_set: - self._console_address = os.path.join(self._temp_dir, - self._name + "-console.sock") - chardev = ('socket,id=console,path=%s,server,nowait' % - self._console_address) - args.extend(['-chardev', chardev]) - if self._console_device_type is None: - args.extend(['-serial', 'chardev:console']) - else: - device = '%s,chardev=console' % self._console_device_type - args.extend(['-device', device]) - return args - - def _pre_launch(self): - self._temp_dir = tempfile.mkdtemp(dir=self._test_dir) - if self._monitor_address is not None: - self._vm_monitor = self._monitor_address - else: - self._vm_monitor = os.path.join(self._temp_dir, - self._name + "-monitor.sock") - self._qemu_log_path = os.path.join(self._temp_dir, self._name + ".log") - self._qemu_log_file = open(self._qemu_log_path, 'wb') - - self._qmp = qmp.QEMUMonitorProtocol(self._vm_monitor, - server=True) - - def _post_launch(self): - self._qmp.accept() - - def _post_shutdown(self): - if self._qemu_log_file is not None: - self._qemu_log_file.close() - self._qemu_log_file = None - - self._qemu_log_path = None - - if self._console_socket is not None: - self._console_socket.close() - self._console_socket = None - - if self._temp_dir is not None: - shutil.rmtree(self._temp_dir) - self._temp_dir = None - - def launch(self): - """ - Launch the VM and make sure we cleanup and expose the - command line/output in case of exception - """ - - if self._launched: - raise QEMUMachineError('VM already launched') - - self._iolog = None - self._qemu_full_args = None - try: - self._launch() - self._launched = True - except: - self.shutdown() - - LOG.debug('Error launching VM') - if self._qemu_full_args: - LOG.debug('Command: %r', ' '.join(self._qemu_full_args)) - if self._iolog: - LOG.debug('Output: %r', self._iolog) - raise Exception(self._iolog) - raise - - def _launch(self): - """ - Launch the VM and establish a QMP connection - """ - devnull = open(os.path.devnull, 'rb') - self._pre_launch() - self._qemu_full_args = (self._wrapper + [self._binary] + - self._base_args() + self._args) - LOG.debug('VM launch command: %r', ' '.join(self._qemu_full_args)) - self._popen = subprocess.Popen(self._qemu_full_args, - stdin=devnull, - stdout=self._qemu_log_file, - stderr=subprocess.STDOUT, - shell=False, - close_fds=False) - self._post_launch() - - def wait(self): - """ - Wait for the VM to power off - """ - self._popen.wait() - self._qmp.close() - self._load_io_log() - self._post_shutdown() - - def shutdown(self): - """ - Terminate the VM and clean up - """ - if self.is_running(): - try: - self._qmp.cmd('quit') - self._qmp.close() - except: - self._popen.kill() - self._popen.wait() - - self._load_io_log() - self._post_shutdown() - - exitcode = self.exitcode() - if exitcode is not None and exitcode < 0: - msg = 'qemu received signal %i: %s' - if self._qemu_full_args: - command = ' '.join(self._qemu_full_args) - else: - command = '' - LOG.warn(msg, -exitcode, command) - - self._launched = False - - def qmp(self, cmd, conv_keys=True, **args): - """ - Invoke a QMP command and return the response dict - """ - qmp_args = dict() - for key, value in args.items(): - if conv_keys: - qmp_args[key.replace('_', '-')] = value - else: - qmp_args[key] = value - - return self._qmp.cmd(cmd, args=qmp_args) - - def command(self, cmd, conv_keys=True, **args): - """ - Invoke a QMP command. - On success return the response dict. - On failure raise an exception. - """ - reply = self.qmp(cmd, conv_keys, **args) - if reply is None: - raise qmp.QMPError("Monitor is closed") - if "error" in reply: - raise MonitorResponseError(reply) - return reply["return"] - - def get_qmp_event(self, wait=False): - """ - Poll for one queued QMP events and return it - """ - if len(self._events) > 0: - return self._events.pop(0) - return self._qmp.pull_event(wait=wait) - - def get_qmp_events(self, wait=False): - """ - Poll for queued QMP events and return a list of dicts - """ - events = self._qmp.get_events(wait=wait) - events.extend(self._events) - del self._events[:] - self._qmp.clear_events() - return events - - @staticmethod - def event_match(event, match=None): - """ - Check if an event matches optional match criteria. - - The match criteria takes the form of a matching subdict. The event is - checked to be a superset of the subdict, recursively, with matching - values whenever the subdict values are not None. - - This has a limitation that you cannot explicitly check for None values. - - Examples, with the subdict queries on the left: - - None matches any object. - - {"foo": None} matches {"foo": {"bar": 1}} - - {"foo": None} matches {"foo": 5} - - {"foo": {"abc": None}} does not match {"foo": {"bar": 1}} - - {"foo": {"rab": 2}} matches {"foo": {"bar": 1, "rab": 2}} - """ - if match is None: - return True - - try: - for key in match: - if key in event: - if not QEMUMachine.event_match(event[key], match[key]): - return False - else: - return False - return True - except TypeError: - # either match or event wasn't iterable (not a dict) - return match == event - - def event_wait(self, name, timeout=60.0, match=None): - """ - event_wait waits for and returns a named event from QMP with a timeout. - - name: The event to wait for. - timeout: QEMUMonitorProtocol.pull_event timeout parameter. - match: Optional match criteria. See event_match for details. - """ - return self.events_wait([(name, match)], timeout) - - def events_wait(self, events, timeout=60.0): - """ - events_wait waits for and returns a named event from QMP with a timeout. - - events: a sequence of (name, match_criteria) tuples. - The match criteria are optional and may be None. - See event_match for details. - timeout: QEMUMonitorProtocol.pull_event timeout parameter. - """ - - def _match(event): - for name, match in events: - if (event['event'] == name and - self.event_match(event, match)): - return True - return False - - # Search cached events - for event in self._events: - if _match(event): - self._events.remove(event) - return event - - # Poll for new events - while True: - event = self._qmp.pull_event(wait=timeout) - if _match(event): - return event - self._events.append(event) - - return None - - def get_log(self): - """ - After self.shutdown or failed qemu execution, this returns the output - of the qemu process. - """ - return self._iolog - - def add_args(self, *args): - """ - Adds to the list of extra arguments to be given to the QEMU binary - """ - self._args.extend(args) - - def set_machine(self, machine_type): - """ - Sets the machine type - - If set, the machine type will be added to the base arguments - of the resulting QEMU command line. - """ - self._machine = machine_type - - def set_console(self, device_type=None): - """ - Sets the device type for a console device - - If set, the console device and a backing character device will - be added to the base arguments of the resulting QEMU command - line. - - This is a convenience method that will either use the provided - device type, or default to a "-serial chardev:console" command - line argument. - - The actual setting of command line arguments will be be done at - machine launch time, as it depends on the temporary directory - to be created. - - @param device_type: the device type, such as "isa-serial". If - None is given (the default value) a "-serial - chardev:console" command line argument will - be used instead, resorting to the machine's - default device type. - """ - self._console_set = True - self._console_device_type = device_type - - @property - def console_socket(self): - """ - Returns a socket connected to the console - """ - if self._console_socket is None: - self._console_socket = socket.socket(socket.AF_UNIX, - socket.SOCK_STREAM) - self._console_socket.connect(self._console_address) - return self._console_socket diff --git a/ucloud/host/qmp/qmp.py b/ucloud/host/qmp/qmp.py deleted file mode 100755 index bf35d71..0000000 --- a/ucloud/host/qmp/qmp.py +++ /dev/null @@ -1,255 +0,0 @@ -# QEMU Monitor Protocol Python class -# -# Copyright (C) 2009, 2010 Red Hat Inc. -# -# Authors: -# Luiz Capitulino -# -# This work is licensed under the terms of the GNU GPL, version 2. See -# the COPYING file in the top-level directory. - -import errno -import json -import logging -import socket - - -class QMPError(Exception): - pass - - -class QMPConnectError(QMPError): - pass - - -class QMPCapabilitiesError(QMPError): - pass - - -class QMPTimeoutError(QMPError): - pass - - -class QEMUMonitorProtocol(object): - #: Logger object for debugging messages - logger = logging.getLogger('QMP') - #: Socket's error class - error = socket.error - #: Socket's timeout - timeout = socket.timeout - - def __init__(self, address, server=False): - """ - Create a QEMUMonitorProtocol class. - - @param address: QEMU address, can be either a unix socket path (string) - or a tuple in the form ( address, port ) for a TCP - connection - @param server: server mode listens on the socket (bool) - @raise socket.error on socket connection errors - @note No connection is established, this is done by the connect() or - accept() methods - """ - self.__events = [] - self.__address = address - self.__sock = self.__get_sock() - self.__sockfile = None - if server: - self.__sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) - self.__sock.bind(self.__address) - self.__sock.listen(1) - - def __get_sock(self): - if isinstance(self.__address, tuple): - family = socket.AF_INET - else: - family = socket.AF_UNIX - return socket.socket(family, socket.SOCK_STREAM) - - def __negotiate_capabilities(self): - greeting = self.__json_read() - if greeting is None or "QMP" not in greeting: - raise QMPConnectError - # Greeting seems ok, negotiate capabilities - resp = self.cmd('qmp_capabilities') - if "return" in resp: - return greeting - raise QMPCapabilitiesError - - def __json_read(self, only_event=False): - while True: - data = self.__sockfile.readline() - if not data: - return - resp = json.loads(data) - if 'event' in resp: - self.logger.debug("<<< %s", resp) - self.__events.append(resp) - if not only_event: - continue - return resp - - def __get_events(self, wait=False): - """ - Check for new events in the stream and cache them in __events. - - @param wait (bool): block until an event is available. - @param wait (float): If wait is a float, treat it as a timeout value. - - @raise QMPTimeoutError: If a timeout float is provided and the timeout - period elapses. - @raise QMPConnectError: If wait is True but no events could be - retrieved or if some other error occurred. - """ - - # Check for new events regardless and pull them into the cache: - self.__sock.setblocking(0) - try: - self.__json_read() - except socket.error as err: - if err[0] == errno.EAGAIN: - # No data available - pass - self.__sock.setblocking(1) - - # Wait for new events, if needed. - # if wait is 0.0, this means "no wait" and is also implicitly false. - if not self.__events and wait: - if isinstance(wait, float): - self.__sock.settimeout(wait) - try: - ret = self.__json_read(only_event=True) - except socket.timeout: - raise QMPTimeoutError("Timeout waiting for event") - except: - raise QMPConnectError("Error while reading from socket") - if ret is None: - raise QMPConnectError("Error while reading from socket") - self.__sock.settimeout(None) - - def connect(self, negotiate=True): - """ - Connect to the QMP Monitor and perform capabilities negotiation. - - @return QMP greeting dict - @raise socket.error on socket connection errors - @raise QMPConnectError if the greeting is not received - @raise QMPCapabilitiesError if fails to negotiate capabilities - """ - self.__sock.connect(self.__address) - self.__sockfile = self.__sock.makefile() - if negotiate: - return self.__negotiate_capabilities() - - def accept(self): - """ - Await connection from QMP Monitor and perform capabilities negotiation. - - @return QMP greeting dict - @raise socket.error on socket connection errors - @raise QMPConnectError if the greeting is not received - @raise QMPCapabilitiesError if fails to negotiate capabilities - """ - self.__sock.settimeout(15) - self.__sock, _ = self.__sock.accept() - self.__sockfile = self.__sock.makefile() - return self.__negotiate_capabilities() - - def cmd_obj(self, qmp_cmd): - """ - Send a QMP command to the QMP Monitor. - - @param qmp_cmd: QMP command to be sent as a Python dict - @return QMP response as a Python dict or None if the connection has - been closed - """ - self.logger.debug(">>> %s", qmp_cmd) - try: - self.__sock.sendall(json.dumps(qmp_cmd).encode('utf-8')) - except socket.error as err: - if err[0] == errno.EPIPE: - return - raise socket.error(err) - resp = self.__json_read() - self.logger.debug("<<< %s", resp) - return resp - - def cmd(self, name, args=None, cmd_id=None): - """ - Build a QMP command and send it to the QMP Monitor. - - @param name: command name (string) - @param args: command arguments (dict) - @param cmd_id: command id (dict, list, string or int) - """ - qmp_cmd = {'execute': name} - if args: - qmp_cmd['arguments'] = args - if cmd_id: - qmp_cmd['id'] = cmd_id - return self.cmd_obj(qmp_cmd) - - def command(self, cmd, **kwds): - """ - Build and send a QMP command to the monitor, report errors if any - """ - ret = self.cmd(cmd, kwds) - if "error" in ret: - raise Exception(ret['error']['desc']) - return ret['return'] - - def pull_event(self, wait=False): - """ - Pulls a single event. - - @param wait (bool): block until an event is available. - @param wait (float): If wait is a float, treat it as a timeout value. - - @raise QMPTimeoutError: If a timeout float is provided and the timeout - period elapses. - @raise QMPConnectError: If wait is True but no events could be - retrieved or if some other error occurred. - - @return The first available QMP event, or None. - """ - self.__get_events(wait) - - if self.__events: - return self.__events.pop(0) - return None - - def get_events(self, wait=False): - """ - Get a list of available QMP events. - - @param wait (bool): block until an event is available. - @param wait (float): If wait is a float, treat it as a timeout value. - - @raise QMPTimeoutError: If a timeout float is provided and the timeout - period elapses. - @raise QMPConnectError: If wait is True but no events could be - retrieved or if some other error occurred. - - @return The list of available QMP events. - """ - self.__get_events(wait) - return self.__events - - def clear_events(self): - """ - Clear current list of pending events. - """ - self.__events = [] - - def close(self): - self.__sock.close() - self.__sockfile.close() - - def settimeout(self, timeout): - self.__sock.settimeout(timeout) - - def get_sock_fd(self): - return self.__sock.fileno() - - def is_scm_available(self): - return self.__sock.family == socket.AF_UNIX diff --git a/ucloud/host/virtualmachine.py b/ucloud/host/virtualmachine.py deleted file mode 100755 index 7524083..0000000 --- a/ucloud/host/virtualmachine.py +++ /dev/null @@ -1,384 +0,0 @@ -# QEMU Manual -# https://qemu.weilnetz.de/doc/qemu-doc.html - -# For QEMU Monitor Protocol Commands Information, See -# https://qemu.weilnetz.de/doc/qemu-doc.html#pcsys_005fmonitor - -import os -import random -import subprocess as sp -import tempfile -import time - -from functools import wraps -from string import Template -from typing import Union -from os.path import join as join_path - -import bitmath -import sshtunnel - -from ucloud.common.helpers import get_ipv6_address -from ucloud.common.request import RequestEntry, RequestType -from ucloud.common.vm import VMEntry, VMStatus -from ucloud.config import (etcd_client, request_pool, - running_vms, vm_pool, env_vars, - image_storage_handler) -from . import qmp -from ucloud.host import logger - - -class VM: - def __init__(self, key, handle, vnc_socket_file): - self.key = key # type: str - self.handle = handle # type: qmp.QEMUMachine - self.vnc_socket_file = vnc_socket_file # type: tempfile.NamedTemporaryFile - - def __repr__(self): - return "VM({})".format(self.key) - - -def delete_network_interface(iface): - try: - sp.check_output(['ip', 'link', 'del', iface]) - except Exception: - pass - - -def resolve_network(network_name, network_owner): - network = etcd_client.get(join_path(env_vars.get("NETWORK_PREFIX"), - network_owner, - network_name), - value_in_json=True) - return network - - -def delete_vm_network(vm_entry): - try: - for network in vm_entry.network: - network_name = network[0] - tap_mac = network[1] - tap_id = network[2] - - delete_network_interface('tap{}'.format(tap_id)) - - owners_vms = vm_pool.by_owner(vm_entry.owner) - owners_running_vms = vm_pool.by_status(VMStatus.running, - _vms=owners_vms) - - networks = map(lambda n: n[0], - map(lambda vm: vm.network, owners_running_vms) - ) - networks_in_use_by_user_vms = [vm[0] for vm in networks] - if network_name not in networks_in_use_by_user_vms: - network_entry = resolve_network(network[0], vm_entry.owner) - if network_entry: - network_type = network_entry.value["type"] - network_id = network_entry.value["id"] - if network_type == "vxlan": - delete_network_interface('br{}'.format(network_id)) - delete_network_interface('vxlan{}'.format(network_id)) - except Exception: - logger.exception("Exception in network interface deletion") - - -def create_dev(script, _id, dev, ip=None): - command = [script, _id, dev] - if ip: - command.append(ip) - try: - output = sp.check_output(command, stderr=sp.PIPE) - except Exception as e: - print(e.stderr) - return None - else: - return output.decode("utf-8").strip() - - -def create_vxlan_br_tap(_id, _dev, tap_id, ip=None): - network_script_base = os.path.join(os.path.dirname(os.path.dirname(__file__)), 'network') - vxlan = create_dev(script=os.path.join(network_script_base, 'create-vxlan.sh'), - _id=_id, dev=_dev) - if vxlan: - bridge = create_dev(script=os.path.join(network_script_base, 'create-bridge.sh'), - _id=_id, dev=vxlan, ip=ip) - if bridge: - tap = create_dev(script=os.path.join(network_script_base, 'create-tap.sh'), - _id=str(tap_id), dev=bridge) - if tap: - return tap - - -def random_bytes(num=6): - return [random.randrange(256) for _ in range(num)] - - -def generate_mac(uaa=False, multicast=False, oui=None, separator=':', byte_fmt='%02x'): - mac = random_bytes() - if oui: - if type(oui) == str: - oui = [int(chunk) for chunk in oui.split(separator)] - mac = oui + random_bytes(num=6 - len(oui)) - else: - if multicast: - mac[0] |= 1 # set bit 0 - else: - mac[0] &= ~1 # clear bit 0 - if uaa: - mac[0] &= ~(1 << 1) # clear bit 1 - else: - mac[0] |= 1 << 1 # set bit 1 - return separator.join(byte_fmt % b for b in mac) - - -def update_radvd_conf(etcd_client): - network_script_base = os.path.join(os.path.dirname(os.path.dirname(__file__)), 'network') - - networks = { - net.value['ipv6']: net.value['id'] - for net in etcd_client.get_prefix('/v1/network/', value_in_json=True) - if net.value.get('ipv6') - } - radvd_template = open(os.path.join(network_script_base, - 'radvd-template.conf'), 'r').read() - radvd_template = Template(radvd_template) - - content = [radvd_template.safe_substitute(bridge='br{}'.format(networks[net]), - prefix=net) - for net in networks if networks.get(net)] - - with open('/etc/radvd.conf', 'w') as radvd_conf: - radvd_conf.writelines(content) - try: - sp.check_output(['systemctl', 'restart', 'radvd']) - except Exception: - sp.check_output(['service', 'radvd', 'restart']) - - -def get_start_command_args(vm_entry, vnc_sock_filename: str, migration=False, migration_port=None): - threads_per_core = 1 - vm_memory = int(bitmath.parse_string_unsafe(vm_entry.specs["ram"]).to_MB()) - vm_cpus = int(vm_entry.specs["cpu"]) - vm_uuid = vm_entry.uuid - vm_networks = vm_entry.network - - command = "-name {}_{}".format(vm_entry.owner, vm_entry.name) - - command += " -drive file={},format=raw,if=virtio,cache=none".format( - image_storage_handler.qemu_path_string(vm_uuid) - ) - command += " -device virtio-rng-pci -vnc unix:{}".format(vnc_sock_filename) - command += " -m {} -smp cores={},threads={}".format( - vm_memory, vm_cpus, threads_per_core - ) - - if migration: - command += " -incoming tcp:[::]:{}".format(migration_port) - - tap = None - for network_mac_and_tap in vm_networks: - network_name, mac, tap = network_mac_and_tap - - _key = os.path.join(env_vars.get('NETWORK_PREFIX'), vm_entry.owner, network_name) - network = etcd_client.get(_key, value_in_json=True) - network_type = network.value["type"] - network_id = str(network.value["id"]) - network_ipv6 = network.value["ipv6"] - - if network_type == "vxlan": - tap = create_vxlan_br_tap(_id=network_id, - _dev=env_vars.get("VXLAN_PHY_DEV"), - tap_id=tap, - ip=network_ipv6) - update_radvd_conf(etcd_client) - - command += " -netdev tap,id=vmnet{net_id},ifname={tap},script=no,downscript=no" \ - " -device virtio-net-pci,netdev=vmnet{net_id},mac={mac}" \ - .format(tap=tap, net_id=network_id, mac=mac) - - return command.split(" ") - - -def create_vm_object(vm_entry, migration=False, migration_port=None): - # NOTE: If migration suddenly stop working, having different - # VNC unix filename on source and destination host can - # be a possible cause of it. - - # REQUIREMENT: Use Unix Socket instead of TCP Port for VNC - vnc_sock_file = tempfile.NamedTemporaryFile() - - qemu_args = get_start_command_args( - vm_entry=vm_entry, - vnc_sock_filename=vnc_sock_file.name, - migration=migration, - migration_port=migration_port, - ) - qemu_machine = qmp.QEMUMachine("/usr/bin/qemu-system-x86_64", args=qemu_args) - return VM(vm_entry.key, qemu_machine, vnc_sock_file) - - -def get_vm(vm_list: list, vm_key) -> Union[VM, None]: - return next((vm for vm in vm_list if vm.key == vm_key), None) - - -def need_running_vm(func): - @wraps(func) - def wrapper(e): - vm = get_vm(running_vms, e.key) - if vm: - try: - status = vm.handle.command("query-status") - logger.debug("VM Status Check - %s", status) - except Exception as exception: - logger.info("%s failed - VM %s %s", func.__name__, e, exception) - else: - return func(e) - - return None - else: - logger.info("%s failed because VM %s is not running", func.__name__, e.key) - return None - - return wrapper - - -def create(vm_entry: VMEntry): - if image_storage_handler.is_vm_image_exists(vm_entry.uuid): - # File Already exists. No Problem Continue - logger.debug("Image for vm %s exists", vm_entry.uuid) - else: - vm_hdd = int(bitmath.parse_string_unsafe(vm_entry.specs["os-ssd"]).to_MB()) - if image_storage_handler.make_vm_image(src=vm_entry.image_uuid, dest=vm_entry.uuid): - if not image_storage_handler.resize_vm_image(path=vm_entry.uuid, size=vm_hdd): - vm_entry.status = VMStatus.error - else: - logger.info("New VM Created") - - -def start(vm_entry: VMEntry, destination_host_key=None, migration_port=None): - _vm = get_vm(running_vms, vm_entry.key) - - # VM already running. No need to proceed further. - if _vm: - logger.info("VM %s already running" % vm_entry.uuid) - return - else: - logger.info("Trying to start %s" % vm_entry.uuid) - if destination_host_key: - launch_vm(vm_entry, migration=True, migration_port=migration_port, - destination_host_key=destination_host_key) - else: - create(vm_entry) - launch_vm(vm_entry) - - -@need_running_vm -def stop(vm_entry): - vm = get_vm(running_vms, vm_entry.key) - vm.handle.shutdown() - if not vm.handle.is_running(): - vm_entry.add_log("Shutdown successfully") - vm_entry.declare_stopped() - vm_pool.put(vm_entry) - running_vms.remove(vm) - delete_vm_network(vm_entry) - - -def delete(vm_entry): - logger.info("Deleting VM | %s", vm_entry) - stop(vm_entry) - - if image_storage_handler.is_vm_image_exists(vm_entry.uuid): - r_status = image_storage_handler.delete_vm_image(vm_entry.uuid) - if r_status: - etcd_client.client.delete(vm_entry.key) - else: - etcd_client.client.delete(vm_entry.key) - -def transfer(request_event): - # This function would run on source host i.e host on which the vm - # is running initially. This host would be responsible for transferring - # vm state to destination host. - - _host, _port = request_event.parameters["host"], request_event.parameters["port"] - _uuid = request_event.uuid - _destination = request_event.destination_host_key - vm = get_vm(running_vms, join_path(env_vars.get('VM_PREFIX'), _uuid)) - - if vm: - tunnel = sshtunnel.SSHTunnelForwarder( - _host, - ssh_username=env_vars.get("ssh_username"), - ssh_pkey=env_vars.get("ssh_pkey"), - remote_bind_address=("127.0.0.1", _port), - ssh_proxy_enabled=True, - ssh_proxy=(_host, 22) - ) - try: - tunnel.start() - except sshtunnel.BaseSSHTunnelForwarderError: - logger.exception("Couldn't establish connection to (%s, 22)", _host) - else: - vm.handle.command( - "migrate", uri="tcp:0.0.0.0:{}".format(tunnel.local_bind_port) - ) - - status = vm.handle.command("query-migrate")["status"] - while status not in ["failed", "completed"]: - time.sleep(2) - status = vm.handle.command("query-migrate")["status"] - - with vm_pool.get_put(request_event.uuid) as source_vm: - if status == "failed": - source_vm.add_log("Migration Failed") - elif status == "completed": - # If VM is successfully migrated then shutdown the VM - # on this host and update hostname to destination host key - source_vm.add_log("Successfully migrated") - source_vm.hostname = _destination - running_vms.remove(vm) - vm.handle.shutdown() - source_vm.in_migration = False # VM transfer finished - finally: - tunnel.close() - - -def launch_vm(vm_entry, migration=False, migration_port=None, destination_host_key=None): - logger.info("Starting %s" % vm_entry.key) - - vm = create_vm_object(vm_entry, migration=migration, migration_port=migration_port) - try: - vm.handle.launch() - except Exception: - logger.exception("Error Occured while starting VM") - vm.handle.shutdown() - - if migration: - # We don't care whether MachineError or any other error occurred - pass - else: - # Error during typical launch of a vm - vm.handle.shutdown() - vm_entry.declare_killed() - vm_pool.put(vm_entry) - else: - vm_entry.vnc_socket = vm.vnc_socket_file.name - running_vms.append(vm) - - if migration: - vm_entry.in_migration = True - r = RequestEntry.from_scratch( - type=RequestType.TransferVM, - hostname=vm_entry.hostname, - parameters={"host": get_ipv6_address(), "port": migration_port}, - uuid=vm_entry.uuid, - destination_host_key=destination_host_key, - request_prefix=env_vars.get("REQUEST_PREFIX") - ) - request_pool.put(r) - else: - # Typical launching of a vm - vm_entry.status = VMStatus.running - vm_entry.add_log("Started successfully") - - vm_pool.put(vm_entry) diff --git a/ucloud/imagescanner/main.py b/ucloud/imagescanner/main.py deleted file mode 100755 index 20ce9d5..0000000 --- a/ucloud/imagescanner/main.py +++ /dev/null @@ -1,78 +0,0 @@ -import json -import os -import subprocess - -from os.path import join as join_path -from ucloud.config import etcd_client, env_vars, image_storage_handler -from ucloud.imagescanner import logger - - -def qemu_img_type(path): - qemu_img_info_command = ["qemu-img", "info", "--output", "json", path] - try: - qemu_img_info = subprocess.check_output(qemu_img_info_command) - except Exception as e: - logger.exception(e) - return None - else: - qemu_img_info = json.loads(qemu_img_info.decode("utf-8")) - return qemu_img_info["format"] - - -def main(): - # We want to get images entries that requests images to be created - images = etcd_client.get_prefix(env_vars.get('IMAGE_PREFIX'), value_in_json=True) - images_to_be_created = list(filter(lambda im: im.value['status'] == 'TO_BE_CREATED', images)) - - for image in images_to_be_created: - try: - image_uuid = image.key.split('/')[-1] - image_owner = image.value['owner'] - image_filename = image.value['filename'] - image_store_name = image.value['store_name'] - image_full_path = join_path(env_vars.get('BASE_DIR'), image_owner, image_filename) - - image_stores = etcd_client.get_prefix(env_vars.get('IMAGE_STORE_PREFIX'), value_in_json=True) - user_image_store = next(filter( - lambda s, store_name=image_store_name: s.value["name"] == store_name, - image_stores - )) - - image_store_pool = user_image_store.value['attributes']['pool'] - - except Exception as e: - logger.exception(e) - else: - # At least our basic data is available - qemu_img_convert_command = ["qemu-img", "convert", "-f", "qcow2", - "-O", "raw", image_full_path, "image.raw"] - - if qemu_img_type(image_full_path) == "qcow2": - try: - # Convert .qcow2 to .raw - subprocess.check_output(qemu_img_convert_command) - except Exception as e: - logger.exception(e) - else: - # Import and Protect - r_status = image_storage_handler.import_image(src="image.raw", - dest=image_uuid, - protect=True) - if r_status: - # Everything is successfully done - image.value["status"] = "CREATED" - etcd_client.put(image.key, json.dumps(image.value)) - - else: - # The user provided image is either not found or of invalid format - image.value["status"] = "INVALID_IMAGE" - etcd_client.put(image.key, json.dumps(image.value)) - - try: - os.remove("image.raw") - except Exception: - pass - - -if __name__ == "__main__": - main() diff --git a/ucloud/metadata/main.py b/ucloud/metadata/main.py deleted file mode 100644 index e7cb33b..0000000 --- a/ucloud/metadata/main.py +++ /dev/null @@ -1,91 +0,0 @@ -import os - -from flask import Flask, request -from flask_restful import Resource, Api - -from ucloud.config import etcd_client, env_vars, vm_pool - -app = Flask(__name__) -api = Api(app) - - -def get_vm_entry(mac_addr): - return next(filter(lambda vm: mac_addr in list(zip(*vm.network))[1], vm_pool.vms), None) - - -# https://stackoverflow.com/questions/37140846/how-to-convert-ipv6-link-local-address-to-mac-address-in-python -def ipv62mac(ipv6): - # remove subnet info if given - subnet_index = ipv6.find('/') - if subnet_index != -1: - ipv6 = ipv6[:subnet_index] - - ipv6_parts = ipv6.split(':') - mac_parts = list() - for ipv6_part in ipv6_parts[-4:]: - while len(ipv6_part) < 4: - ipv6_part = '0' + ipv6_part - mac_parts.append(ipv6_part[:2]) - mac_parts.append(ipv6_part[-2:]) - - # modify parts to match MAC value - mac_parts[0] = '%02x' % (int(mac_parts[0], 16) ^ 2) - del mac_parts[4] - del mac_parts[3] - return ':'.join(mac_parts) - - -class Root(Resource): - @staticmethod - def get(): - data = get_vm_entry(ipv62mac(request.remote_addr)) - - if not data: - return {'message': 'Metadata for such VM does not exists.'}, 404 - else: - - # {env_vars.get('USER_PREFIX')}/{realm}/{name}/key - etcd_key = os.path.join(env_vars.get('USER_PREFIX'), data.value['owner_realm'], - data.value['owner'], 'key') - etcd_entry = etcd_client.get_prefix(etcd_key, value_in_json=True) - user_personal_ssh_keys = [key.value for key in etcd_entry] - data.value['metadata']['ssh-keys'] += user_personal_ssh_keys - return data.value['metadata'], 200 - - @staticmethod - def post(): - return {'message': 'Previous Implementation is deprecated.'} - # data = etcd_client.get("/v1/metadata/{}".format(request.remote_addr), value_in_json=True) - # print(data) - # if data: - # for k in request.json: - # if k not in data.value: - # data.value[k] = request.json[k] - # if k.endswith("-list"): - # data.value[k] = [request.json[k]] - # else: - # if k.endswith("-list"): - # data.value[k].append(request.json[k]) - # else: - # data.value[k] = request.json[k] - # etcd_client.put("/v1/metadata/{}".format(request.remote_addr), - # data.value, value_in_json=True) - # else: - # data = {} - # for k in request.json: - # data[k] = request.json[k] - # if k.endswith("-list"): - # data[k] = [request.json[k]] - # etcd_client.put("/v1/metadata/{}".format(request.remote_addr), - # data, value_in_json=True) - - -api.add_resource(Root, '/') - - -def main(): - app.run(debug=True, host="::", port="80") - - -if __name__ == '__main__': - main() diff --git a/ucloud/sanity_checks.py b/ucloud/sanity_checks.py deleted file mode 100644 index 143f767..0000000 --- a/ucloud/sanity_checks.py +++ /dev/null @@ -1,33 +0,0 @@ -import sys -import subprocess as sp - -from os.path import isdir -from ucloud.config import env_vars - - -def check(): - ######################### - # ucloud-image-scanner # - ######################### - if env_vars.get('STORAGE_BACKEND') == 'filesystem' and not isdir(env_vars.get('IMAGE_DIR')): - print("You have set STORAGE_BACKEND to filesystem. So," - "the {} must exists. But, it don't".format(env_vars.get('IMAGE_DIR'))) - sys.exit(1) - - try: - sp.check_output(['which', 'qemu-img']) - except Exception: - print("qemu-img missing") - sys.exit(1) - - ############### - # ucloud-host # - ############### - - if env_vars.get('STORAGE_BACKEND') == 'filesystem' and not isdir(env_vars.get('VM_DIR')): - print("You have set STORAGE_BACKEND to filesystem. So, the vm directory mentioned" - " in .env file must exists. But, it don't.") - sys.exit(1) - -if __name__ == "__main__": - check() \ No newline at end of file diff --git a/ucloud/scheduler/__init__.py b/ucloud/scheduler/__init__.py deleted file mode 100644 index 95e1be0..0000000 --- a/ucloud/scheduler/__init__.py +++ /dev/null @@ -1,3 +0,0 @@ -import logging - -logger = logging.getLogger(__name__) \ No newline at end of file diff --git a/ucloud/scheduler/main.py b/ucloud/scheduler/main.py deleted file mode 100755 index e2c975a..0000000 --- a/ucloud/scheduler/main.py +++ /dev/null @@ -1,93 +0,0 @@ -# TODO -# 1. send an email to an email address defined by env['admin-email'] -# if resources are finished -# 2. Introduce a status endpoint of the scheduler - -# maybe expose a prometheus compatible output - -from ucloud.common.request import RequestEntry, RequestType -from ucloud.config import etcd_client -from ucloud.config import host_pool, request_pool, vm_pool, env_vars -from .helper import (get_suitable_host, dead_host_mitigation, dead_host_detection, - assign_host, NoSuitableHostFound) -from . import logger - - -def main(): - logger.info("%s SESSION STARTED %s", '*' * 5, '*' * 5) - - pending_vms = [] - - for request_iterator in [ - etcd_client.get_prefix(env_vars.get('REQUEST_PREFIX'), value_in_json=True), - etcd_client.watch_prefix(env_vars.get('REQUEST_PREFIX'), timeout=5, value_in_json=True), - ]: - for request_event in request_iterator: - request_entry = RequestEntry(request_event) - # Never Run time critical mechanism inside timeout - # mechanism because timeout mechanism only comes - # when no other event is happening. It means under - # heavy load there would not be a timeout event. - if request_entry.type == "TIMEOUT": - - # Detect hosts that are dead and set their status - # to "DEAD", and their VMs' status to "KILLED" - dead_hosts = dead_host_detection() - if dead_hosts: - logger.debug("Dead hosts: %s", dead_hosts) - dead_host_mitigation(dead_hosts) - - # If there are VMs that weren't assigned a host - # because there wasn't a host available which - # meets requirement of that VM then we would - # create a new ScheduleVM request for that VM - # on our behalf. - while pending_vms: - pending_vm_entry = pending_vms.pop() - r = RequestEntry.from_scratch(type="ScheduleVM", - uuid=pending_vm_entry.uuid, - hostname=pending_vm_entry.hostname, - request_prefix=env_vars.get("REQUEST_PREFIX")) - request_pool.put(r) - - elif request_entry.type == RequestType.ScheduleVM: - logger.debug("%s, %s", request_entry.key, request_entry.value) - - vm_entry = vm_pool.get(request_entry.uuid) - if vm_entry is None: - logger.info("Trying to act on {} but it is deleted".format(request_entry.uuid)) - continue - etcd_client.client.delete(request_entry.key) # consume Request - - # If the Request is about a VM which is labelled as "migration" - # and has a destination - if hasattr(request_entry, "migration") and request_entry.migration \ - and hasattr(request_entry, "destination") and request_entry.destination: - try: - get_suitable_host(vm_specs=vm_entry.specs, - hosts=[host_pool.get(request_entry.destination)]) - except NoSuitableHostFound: - logger.info("Requested destination host doesn't have enough capacity" - "to hold %s" % vm_entry.uuid) - else: - r = RequestEntry.from_scratch(type=RequestType.InitVMMigration, - uuid=request_entry.uuid, - destination=request_entry.destination, - request_prefix=env_vars.get("REQUEST_PREFIX")) - request_pool.put(r) - - # If the Request is about a VM that just want to get started/created - else: - # assign_host only returns None when we couldn't be able to assign - # a host to a VM because of resource constraints - try: - assign_host(vm_entry) - except NoSuitableHostFound: - vm_entry.add_log("Can't schedule VM. No Resource Left.") - vm_pool.put(vm_entry) - - pending_vms.append(vm_entry) - logger.info("No Resource Left. Emailing admin....") - - -if __name__ == "__main__": - main() diff --git a/uncloud/__init__.py b/uncloud/__init__.py new file mode 100644 index 0000000..2920f47 --- /dev/null +++ b/uncloud/__init__.py @@ -0,0 +1,2 @@ +class UncloudException(Exception): + pass diff --git a/ucloud/api/README.md b/uncloud/api/README.md similarity index 100% rename from ucloud/api/README.md rename to uncloud/api/README.md diff --git a/ucloud/api/__init__.py b/uncloud/api/__init__.py similarity index 100% rename from ucloud/api/__init__.py rename to uncloud/api/__init__.py diff --git a/ucloud/api/common_fields.py b/uncloud/api/common_fields.py similarity index 72% rename from ucloud/api/common_fields.py rename to uncloud/api/common_fields.py index e9903ac..ba9fb37 100755 --- a/ucloud/api/common_fields.py +++ b/uncloud/api/common_fields.py @@ -1,6 +1,6 @@ import os -from ucloud.config import etcd_client, env_vars +from uncloud.common.shared import shared class Optional: @@ -19,12 +19,16 @@ class Field: def is_valid(self): if self.value == KeyError: - self.add_error("'{}' field is a required field".format(self.name)) + self.add_error( + "'{}' field is a required field".format(self.name) + ) else: if isinstance(self.value, Optional): pass elif not isinstance(self.value, self.type): - self.add_error("Incorrect Type for '{}' field".format(self.name)) + self.add_error( + "Incorrect Type for '{}' field".format(self.name) + ) else: self.validation() @@ -48,6 +52,8 @@ class VmUUIDField(Field): self.validation = self.vm_uuid_validation def vm_uuid_validation(self): - r = etcd_client.get(os.path.join(env_vars.get('VM_PREFIX'), self.uuid)) + r = shared.etcd_client.get( + os.path.join(shared.settings["etcd"]["vm_prefix"], self.uuid) + ) if not r: self.add_error("VM with uuid {} does not exists".format(self.uuid)) diff --git a/uncloud/api/create_image_store.py b/uncloud/api/create_image_store.py new file mode 100755 index 0000000..90e0f92 --- /dev/null +++ b/uncloud/api/create_image_store.py @@ -0,0 +1,19 @@ +import json +import os + +from uuid import uuid4 + +from uncloud.common.shared import shared + +data = { + 'is_public': True, + 'type': 'ceph', + 'name': 'images', + 'description': 'first ever public image-store', + 'attributes': {'list': [], 'key': [], 'pool': 'images'}, +} + +shared.etcd_client.put( + os.path.join(shared.settings['etcd']['image_store_prefix'], uuid4().hex), + json.dumps(data), +) diff --git a/ucloud/api/helper.py b/uncloud/api/helper.py similarity index 55% rename from ucloud/api/helper.py rename to uncloud/api/helper.py index 63d2f90..8ceb3a6 100755 --- a/ucloud/api/helper.py +++ b/uncloud/api/helper.py @@ -1,48 +1,51 @@ import binascii import ipaddress import random -import subprocess as sp - +import logging import requests from pyotp import TOTP -from ucloud.config import vm_pool, env_vars +from uncloud.common.shared import shared + +logger = logging.getLogger(__name__) def check_otp(name, realm, token): try: data = { - "auth_name": env_vars.get("AUTH_NAME"), - "auth_token": TOTP(env_vars.get("AUTH_SEED")).now(), - "auth_realm": env_vars.get("AUTH_REALM"), + "auth_name": shared.settings["otp"]["auth_name"], + "auth_token": TOTP(shared.settings["otp"]["auth_seed"]).now(), + "auth_realm": shared.settings["otp"]["auth_realm"], "name": name, "realm": realm, "token": token, } - except binascii.Error: + except binascii.Error as err: + logger.error( + "Cannot compute OTP for seed: {}".format( + shared.settings["otp"]["auth_seed"] + ) + ) return 400 response = requests.post( - "{OTP_SERVER}{OTP_VERIFY_ENDPOINT}".format( - OTP_SERVER=env_vars.get("OTP_SERVER", ""), - OTP_VERIFY_ENDPOINT=env_vars.get("OTP_VERIFY_ENDPOINT", "verify/"), - ), - json=data, + shared.settings["otp"]["verification_controller_url"], json=data ) return response.status_code def resolve_vm_name(name, owner): """Return UUID of Virtual Machine of name == name and owner == owner - + Input: name of vm, owner of vm. Output: uuid of vm if found otherwise None """ result = next( filter( - lambda vm: vm.value["owner"] == owner and vm.value["name"] == name, - vm_pool.vms, + lambda vm: vm.value["owner"] == owner + and vm.value["name"] == name, + shared.vm_pool.vms, ), None, ) @@ -54,7 +57,7 @@ def resolve_vm_name(name, owner): def resolve_image_name(name, etcd_client): """Return image uuid given its name and its store - + * If the provided name is not in correct format i.e {store_name}:{image_name} return ValueError * If no such image found then return KeyError @@ -70,26 +73,35 @@ def resolve_image_name(name, etcd_client): """ Examples, where it would work and where it would raise exception "images:alpine" --> ["images", "alpine"] - + "images" --> ["images"] it would raise Exception as non enough value to unpack - + "images:alpine:meow" --> ["images", "alpine", "meow"] it would raise Exception as too many values to unpack """ store_name, image_name = store_name_and_image_name except Exception: - raise ValueError("Image name not in correct format i.e {store_name}:{image_name}") + raise ValueError( + "Image name not in correct format i.e {store_name}:{image_name}" + ) - images = etcd_client.get_prefix(env_vars.get('IMAGE_PREFIX'), value_in_json=True) + images = etcd_client.get_prefix( + shared.settings["etcd"]["image_prefix"], value_in_json=True + ) # Try to find image with name == image_name and store_name == store_name try: - image = next(filter(lambda im: im.value['name'] == image_name - and im.value['store_name'] == store_name, images)) + image = next( + filter( + lambda im: im.value["name"] == image_name + and im.value["store_name"] == store_name, + images, + ) + ) except StopIteration: raise KeyError("No image with name {} found.".format(name)) else: - image_uuid = image.key.split('/')[-1] + image_uuid = image.key.split("/")[-1] return image_uuid @@ -98,7 +110,7 @@ def random_bytes(num=6): return [random.randrange(256) for _ in range(num)] -def generate_mac(uaa=False, multicast=False, oui=None, separator=':', byte_fmt='%02x'): +def generate_mac(uaa=False, multicast=False, oui=None, separator=":", byte_fmt="%02x"): mac = random_bytes() if oui: if type(oui) == str: @@ -116,36 +128,6 @@ def generate_mac(uaa=False, multicast=False, oui=None, separator=':', byte_fmt=' return separator.join(byte_fmt % b for b in mac) -def get_ip_addr(mac_address, device): - """Return IP address of a device provided its mac address / link local address - and the device with which it is connected. - - For Example, if we call get_ip_addr(mac_address="52:54:00:12:34:56", device="br0") - the following two scenarios can happen - 1. It would return None if we can't be able to find device whose mac_address is equal - to the arg:mac_address or the mentioned arg:device does not exists or the ip address - we found is local. - 2. It would return ip_address of device whose mac_address is equal to arg:mac_address - and is connected/neighbor of arg:device - """ - try: - output = sp.check_output(['ip', '-6', 'neigh', 'show', 'dev', device], stderr=sp.PIPE) - except sp.CalledProcessError: - return None - else: - result = [] - output = output.strip().decode("utf-8") - output = output.split("\n") - for entry in output: - entry = entry.split() - if entry: - ip = ipaddress.ip_address(entry[0]) - mac = entry[2] - if ip.is_global and mac_address == mac: - result.append(ip) - return result - - def mac2ipv6(mac, prefix): # only accept MACs separated by a colon parts = mac.split(":") @@ -158,8 +140,9 @@ def mac2ipv6(mac, prefix): # format output ipv6_parts = [str(0)] * 4 for i in range(0, len(parts), 2): - ipv6_parts.append("".join(parts[i:i + 2])) + ipv6_parts.append("".join(parts[i : i + 2])) lower_part = ipaddress.IPv6Address(":".join(ipv6_parts)) prefix = ipaddress.IPv6Address(prefix) return str(prefix + int(lower_part)) + diff --git a/uncloud/api/main.py b/uncloud/api/main.py new file mode 100644 index 0000000..73e8e21 --- /dev/null +++ b/uncloud/api/main.py @@ -0,0 +1,600 @@ +import json +import pynetbox +import logging +import argparse + +from uuid import uuid4 +from os.path import join as join_path + +from flask import Flask, request +from flask_restful import Resource, Api +from werkzeug.exceptions import HTTPException + +from uncloud.common.shared import shared + +from uncloud.common import counters +from uncloud.common.vm import VMStatus +from uncloud.common.request import RequestEntry, RequestType +from uncloud.api import schemas +from uncloud.api.helper import generate_mac, mac2ipv6 +from uncloud import UncloudException + +logger = logging.getLogger(__name__) + +app = Flask(__name__) +api = Api(app) +app.logger.handlers.clear() + +arg_parser = argparse.ArgumentParser('api', add_help=False) +arg_parser.add_argument('--port', '-p') + + +@app.errorhandler(Exception) +def handle_exception(e): + app.logger.error(e) + # pass through HTTP errors + if isinstance(e, HTTPException): + return e + + # now you're handling non-HTTP exceptions only + return {'message': 'Server Error'}, 500 + + +class CreateVM(Resource): + """API Request to Handle Creation of VM""" + + @staticmethod + def post(): + data = request.json + validator = schemas.CreateVMSchema(data) + if validator.is_valid(): + vm_uuid = uuid4().hex + vm_key = join_path(shared.settings['etcd']['vm_prefix'], vm_uuid) + specs = { + 'cpu': validator.specs['cpu'], + 'ram': validator.specs['ram'], + 'os-ssd': validator.specs['os-ssd'], + 'hdd': validator.specs['hdd'], + } + macs = [generate_mac() for _ in range(len(data['network']))] + tap_ids = [ + counters.increment_etcd_counter( + shared.etcd_client, shared.settings['etcd']['tap_counter'] + ) + for _ in range(len(data['network'])) + ] + vm_entry = { + 'name': data['vm_name'], + 'owner': data['name'], + 'owner_realm': data['realm'], + 'specs': specs, + 'hostname': '', + 'status': VMStatus.stopped, + 'image_uuid': validator.image_uuid, + 'log': [], + 'vnc_socket': '', + 'network': list(zip(data['network'], macs, tap_ids)), + 'metadata': {'ssh-keys': []}, + 'in_migration': False, + } + shared.etcd_client.put(vm_key, vm_entry, value_in_json=True) + + # Create ScheduleVM Request + r = RequestEntry.from_scratch( + type=RequestType.ScheduleVM, + uuid=vm_uuid, + request_prefix=shared.settings['etcd']['request_prefix'], + ) + shared.request_pool.put(r) + + return {'message': 'VM Creation Queued'}, 200 + return validator.get_errors(), 400 + + +class VmStatus(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.VMStatusSchema(data) + if validator.is_valid(): + vm = shared.vm_pool.get( + join_path(shared.settings['etcd']['vm_prefix'], data['uuid']) + ) + vm_value = vm.value.copy() + vm_value['ip'] = [] + for network_mac_and_tap in vm.network: + network_name, mac, tap = network_mac_and_tap + network = shared.etcd_client.get( + join_path( + shared.settings['etcd']['network_prefix'], + data['name'], + network_name, + ), + value_in_json=True, + ) + ipv6_addr = ( + network.value.get('ipv6').split('::')[0] + '::' + ) + vm_value['ip'].append(mac2ipv6(mac, ipv6_addr)) + vm.value = vm_value + return vm.value + else: + return validator.get_errors(), 400 + + +class CreateImage(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.CreateImageSchema(data) + if validator.is_valid(): + file_entry = shared.etcd_client.get( + join_path(shared.settings['etcd']['file_prefix'], data['uuid']) + ) + file_entry_value = json.loads(file_entry.value) + + image_entry_json = { + 'status': 'TO_BE_CREATED', + 'owner': file_entry_value['owner'], + 'filename': file_entry_value['filename'], + 'name': data['name'], + 'store_name': data['image_store'], + 'visibility': 'public', + } + shared.etcd_client.put( + join_path( + shared.settings['etcd']['image_prefix'], data['uuid'] + ), + json.dumps(image_entry_json), + ) + + return {'message': 'Image queued for creation.'} + return validator.get_errors(), 400 + + +class ListPublicImages(Resource): + @staticmethod + def get(): + images = shared.etcd_client.get_prefix( + shared.settings['etcd']['image_prefix'], value_in_json=True + ) + r = {'images': []} + for image in images: + image_key = '{}:{}'.format( + image.value['store_name'], image.value['name'] + ) + r['images'].append( + {'name': image_key, 'status': image.value['status']} + ) + return r, 200 + + +class VMAction(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.VmActionSchema(data) + + if validator.is_valid(): + vm_entry = shared.vm_pool.get( + join_path(shared.settings['etcd']['vm_prefix'], data['uuid']) + ) + action = data['action'] + + if action == 'start': + action = 'schedule' + + if action == 'delete' and vm_entry.hostname == '': + if shared.storage_handler.is_vm_image_exists( + vm_entry.uuid + ): + r_status = shared.storage_handler.delete_vm_image( + vm_entry.uuid + ) + if r_status: + shared.etcd_client.client.delete(vm_entry.key) + return {'message': 'VM successfully deleted'} + else: + logger.error( + 'Some Error Occurred while deleting VM' + ) + return {'message': 'VM deletion unsuccessfull'} + else: + shared.etcd_client.client.delete(vm_entry.key) + return {'message': 'VM successfully deleted'} + + r = RequestEntry.from_scratch( + type='{}VM'.format(action.title()), + uuid=data['uuid'], + hostname=vm_entry.hostname, + request_prefix=shared.settings['etcd']['request_prefix'], + ) + shared.request_pool.put(r) + return ( + {'message': 'VM {} Queued'.format(action.title())}, + 200, + ) + else: + return validator.get_errors(), 400 + + +class VMMigration(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.VmMigrationSchema(data) + + if validator.is_valid(): + vm = shared.vm_pool.get(data['uuid']) + r = RequestEntry.from_scratch( + type=RequestType.InitVMMigration, + uuid=vm.uuid, + hostname=join_path( + shared.settings['etcd']['host_prefix'], + validator.destination.value, + ), + request_prefix=shared.settings['etcd']['request_prefix'], + ) + + shared.request_pool.put(r) + return ( + {'message': 'VM Migration Initialization Queued'}, + 200, + ) + else: + return validator.get_errors(), 400 + + +class ListUserVM(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.OTPSchema(data) + + if validator.is_valid(): + vms = shared.etcd_client.get_prefix( + shared.settings['etcd']['vm_prefix'], value_in_json=True + ) + return_vms = [] + user_vms = filter( + lambda v: v.value['owner'] == data['name'], vms + ) + for vm in user_vms: + return_vms.append( + { + 'name': vm.value['name'], + 'vm_uuid': vm.key.split('/')[-1], + 'specs': vm.value['specs'], + 'status': vm.value['status'], + 'hostname': vm.value['hostname'], + 'vnc_socket': vm.value.get('vnc_socket', None), + } + ) + if return_vms: + return {'message': return_vms}, 200 + return {'message': 'No VM found'}, 404 + + else: + return validator.get_errors(), 400 + + +class ListUserFiles(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.OTPSchema(data) + + if validator.is_valid(): + files = shared.etcd_client.get_prefix( + shared.settings['etcd']['file_prefix'], value_in_json=True + ) + return_files = [] + user_files = [f for f in files if f.value['owner'] == data['name']] + for file in user_files: + file_uuid = file.key.split('/')[-1] + file = file.value + file['uuid'] = file_uuid + + file.pop('sha512sum', None) + file.pop('owner', None) + + return_files.append(file) + return {'message': return_files}, 200 + else: + return validator.get_errors(), 400 + + +class CreateHost(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.CreateHostSchema(data) + if validator.is_valid(): + host_key = join_path( + shared.settings['etcd']['host_prefix'], uuid4().hex + ) + host_entry = { + 'specs': data['specs'], + 'hostname': data['hostname'], + 'status': 'DEAD', + 'last_heartbeat': '', + } + shared.etcd_client.put( + host_key, host_entry, value_in_json=True + ) + + return {'message': 'Host Created'}, 200 + + return validator.get_errors(), 400 + + +class ListHost(Resource): + @staticmethod + def get(): + hosts = shared.host_pool.hosts + r = { + host.key: { + 'status': host.status, + 'specs': host.specs, + 'hostname': host.hostname, + } + for host in hosts + } + return r, 200 + + +class GetSSHKeys(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.GetSSHSchema(data) + if validator.is_valid(): + if not validator.key_name.value: + + # {user_prefix}/{realm}/{name}/key/ + etcd_key = join_path( + shared.settings['etcd']['user_prefix'], + data['realm'], + data['name'], + 'key', + ) + etcd_entry = shared.etcd_client.get_prefix( + etcd_key, value_in_json=True + ) + + keys = { + key.key.split('/')[-1]: key.value + for key in etcd_entry + } + return {'keys': keys} + else: + + # {user_prefix}/{realm}/{name}/key/{key_name} + etcd_key = join_path( + shared.settings['etcd']['user_prefix'], + data['realm'], + data['name'], + 'key', + data['key_name'], + ) + etcd_entry = shared.etcd_client.get( + etcd_key, value_in_json=True + ) + + if etcd_entry: + return { + 'keys': { + etcd_entry.key.split('/')[ + -1 + ]: etcd_entry.value + } + } + else: + return {'keys': {}} + else: + return validator.get_errors(), 400 + + +class AddSSHKey(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.AddSSHSchema(data) + if validator.is_valid(): + + # {user_prefix}/{realm}/{name}/key/{key_name} + etcd_key = join_path( + shared.settings['etcd']['user_prefix'], + data['realm'], + data['name'], + 'key', + data['key_name'], + ) + etcd_entry = shared.etcd_client.get( + etcd_key, value_in_json=True + ) + if etcd_entry: + return { + 'message': 'Key with name "{}" already exists'.format( + data['key_name'] + ) + } + else: + # Key Not Found. It implies user' haven't added any key yet. + shared.etcd_client.put( + etcd_key, data['key'], value_in_json=True + ) + return {'message': 'Key added successfully'} + else: + return validator.get_errors(), 400 + + +class RemoveSSHKey(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.RemoveSSHSchema(data) + if validator.is_valid(): + + # {user_prefix}/{realm}/{name}/key/{key_name} + etcd_key = join_path( + shared.settings['etcd']['user_prefix'], + data['realm'], + data['name'], + 'key', + data['key_name'], + ) + etcd_entry = shared.etcd_client.get( + etcd_key, value_in_json=True + ) + if etcd_entry: + shared.etcd_client.client.delete(etcd_key) + return {'message': 'Key successfully removed.'} + else: + return { + 'message': 'No Key with name "{}" Exists at all.'.format( + data['key_name'] + ) + } + else: + return validator.get_errors(), 400 + + +class CreateNetwork(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.CreateNetwork(data) + + if validator.is_valid(): + + network_entry = { + 'id': counters.increment_etcd_counter( + shared.etcd_client, shared.settings['etcd']['vxlan_counter'] + ), + 'type': data['type'], + } + if validator.user.value: + try: + nb = pynetbox.api( + url=shared.settings['netbox']['url'], + token=shared.settings['netbox']['token'], + ) + nb_prefix = nb.ipam.prefixes.get( + prefix=shared.settings['network']['prefix'] + ) + prefix = nb_prefix.available_prefixes.create( + data={ + 'prefix_length': int( + shared.settings['network']['prefix_length'] + ), + 'description': '{}\'s network "{}"'.format( + data['name'], data['network_name'] + ), + 'is_pool': True, + } + ) + except Exception as err: + app.logger.error(err) + return { + 'message': 'Error occured while creating network.' + } + else: + network_entry['ipv6'] = prefix['prefix'] + else: + network_entry['ipv6'] = 'fd00::/64' + + network_key = join_path( + shared.settings['etcd']['network_prefix'], + data['name'], + data['network_name'], + ) + shared.etcd_client.put( + network_key, network_entry, value_in_json=True + ) + return {'message': 'Network successfully added.'} + else: + return validator.get_errors(), 400 + + +class ListUserNetwork(Resource): + @staticmethod + def post(): + data = request.json + validator = schemas.OTPSchema(data) + + if validator.is_valid(): + prefix = join_path( + shared.settings['etcd']['network_prefix'], data['name'] + ) + networks = shared.etcd_client.get_prefix( + prefix, value_in_json=True + ) + user_networks = [] + for net in networks: + net.value['name'] = net.key.split('/')[-1] + user_networks.append(net.value) + return {'networks': user_networks}, 200 + else: + return validator.get_errors(), 400 + + +api.add_resource(CreateVM, '/vm/create') +api.add_resource(VmStatus, '/vm/status') + +api.add_resource(VMAction, '/vm/action') +api.add_resource(VMMigration, '/vm/migrate') + +api.add_resource(CreateImage, '/image/create') +api.add_resource(ListPublicImages, '/image/list-public') + +api.add_resource(ListUserVM, '/user/vms') +api.add_resource(ListUserFiles, '/user/files') +api.add_resource(ListUserNetwork, '/user/networks') + +api.add_resource(AddSSHKey, '/user/add-ssh') +api.add_resource(RemoveSSHKey, '/user/remove-ssh') +api.add_resource(GetSSHKeys, '/user/get-ssh') + +api.add_resource(CreateHost, '/host/create') +api.add_resource(ListHost, '/host/list') + +api.add_resource(CreateNetwork, '/network/create') + + +def main(arguments): + debug = arguments['debug'] + port = arguments['port'] + + try: + image_stores = list( + shared.etcd_client.get_prefix( + shared.settings['etcd']['image_store_prefix'], value_in_json=True + ) + ) + except KeyError: + image_stores = False + + # Do not inject default values that might be very wrong + # fail when required, not before + # + # if not image_stores: + # data = { + # 'is_public': True, + # 'type': 'ceph', + # 'name': 'images', + # 'description': 'first ever public image-store', + # 'attributes': {'list': [], 'key': [], 'pool': 'images'}, + # } + + # shared.etcd_client.put( + # join_path( + # shared.settings['etcd']['image_store_prefix'], uuid4().hex + # ), + # json.dumps(data), + # ) + + try: + app.run(host='::', port=port, debug=debug) + except OSError as e: + raise UncloudException('Failed to start Flask: {}'.format(e)) diff --git a/ucloud/api/schemas.py b/uncloud/api/schemas.py similarity index 58% rename from ucloud/api/schemas.py rename to uncloud/api/schemas.py index c4f60ca..87f20c9 100755 --- a/ucloud/api/schemas.py +++ b/uncloud/api/schemas.py @@ -1,6 +1,6 @@ """ This module contain classes thats validates and intercept/modify -data coming from ucloud-cli (user) +data coming from uncloud-cli (user) It was primarily developed as an alternative to argument parser of Flask_Restful which is going to be deprecated. I also tried @@ -19,10 +19,10 @@ import os import bitmath -from ucloud.common.host import HostStatus -from ucloud.common.vm import VMStatus -from ucloud.config import etcd_client, env_vars, vm_pool, host_pool -from . import helper +from uncloud.common.host import HostStatus +from uncloud.common.vm import VMStatus +from uncloud.common.shared import shared +from . import helper, logger from .common_fields import Field, VmUUIDField from .helper import check_otp, resolve_vm_name @@ -79,7 +79,12 @@ class OTPSchema(BaseSchema): super().__init__(data=data, fields=_fields) def validation(self): - if check_otp(self.name.value, self.realm.value, self.token.value) != 200: + if ( + check_otp( + self.name.value, self.realm.value, self.token.value + ) + != 200 + ): self.add_error("Wrong Credentials") @@ -91,7 +96,9 @@ class CreateImageSchema(BaseSchema): # Fields self.uuid = Field("uuid", str, data.get("uuid", KeyError)) self.name = Field("name", str, data.get("name", KeyError)) - self.image_store = Field("image_store", str, data.get("image_store", KeyError)) + self.image_store = Field( + "image_store", str, data.get("image_store", KeyError) + ) # Validations self.uuid.validation = self.file_uuid_validation @@ -102,34 +109,51 @@ class CreateImageSchema(BaseSchema): super().__init__(data, fields) def file_uuid_validation(self): - file_entry = etcd_client.get(os.path.join(env_vars.get('FILE_PREFIX'), self.uuid.value)) + file_entry = shared.etcd_client.get( + os.path.join( + shared.shared.shared.shared.shared.settings["etcd"]["file_prefix"], self.uuid.value + ) + ) if file_entry is None: self.add_error( - "Image File with uuid '{}' Not Found".format(self.uuid.value) + "Image File with uuid '{}' Not Found".format( + self.uuid.value + ) ) def image_store_name_validation(self): - image_stores = list(etcd_client.get_prefix(env_vars.get('IMAGE_STORE_PREFIX'))) + image_stores = list( + shared.etcd_client.get_prefix( + shared.shared.shared.shared.shared.settings["etcd"]["image_store_prefix"] + ) + ) image_store = next( filter( - lambda s: json.loads(s.value)["name"] == self.image_store.value, + lambda s: json.loads(s.value)["name"] + == self.image_store.value, image_stores, ), None, ) if not image_store: - self.add_error("Store '{}' does not exists".format(self.image_store.value)) + self.add_error( + "Store '{}' does not exists".format( + self.image_store.value + ) + ) # Host Operations + class CreateHostSchema(OTPSchema): def __init__(self, data): - self.parsed_specs = {} # Fields self.specs = Field("specs", dict, data.get("specs", KeyError)) - self.hostname = Field("hostname", str, data.get("hostname", KeyError)) + self.hostname = Field( + "hostname", str, data.get("hostname", KeyError) + ) # Validation self.specs.validation = self.specs_validation @@ -141,22 +165,28 @@ class CreateHostSchema(OTPSchema): def specs_validation(self): ALLOWED_BASE = 10 - _cpu = self.specs.value.get('cpu', KeyError) - _ram = self.specs.value.get('ram', KeyError) - _os_ssd = self.specs.value.get('os-ssd', KeyError) - _hdd = self.specs.value.get('hdd', KeyError) + _cpu = self.specs.value.get("cpu", KeyError) + _ram = self.specs.value.get("ram", KeyError) + _os_ssd = self.specs.value.get("os-ssd", KeyError) + _hdd = self.specs.value.get("hdd", KeyError) if KeyError in [_cpu, _ram, _os_ssd, _hdd]: - self.add_error("You must specify CPU, RAM and OS-SSD in your specs") + self.add_error( + "You must specify CPU, RAM and OS-SSD in your specs" + ) return None try: parsed_ram = bitmath.parse_string_unsafe(_ram) parsed_os_ssd = bitmath.parse_string_unsafe(_os_ssd) if parsed_ram.base != ALLOWED_BASE: - self.add_error("Your specified RAM is not in correct units") + self.add_error( + "Your specified RAM is not in correct units" + ) if parsed_os_ssd.base != ALLOWED_BASE: - self.add_error("Your specified OS-SSD is not in correct units") + self.add_error( + "Your specified OS-SSD is not in correct units" + ) if _cpu < 1: self.add_error("CPU must be atleast 1") @@ -171,7 +201,9 @@ class CreateHostSchema(OTPSchema): for hdd in _hdd: _parsed_hdd = bitmath.parse_string_unsafe(hdd) if _parsed_hdd.base != ALLOWED_BASE: - self.add_error("Your specified HDD is not in correct units") + self.add_error( + "Your specified HDD is not in correct units" + ) break else: parsed_hdd.append(str(_parsed_hdd)) @@ -182,15 +214,17 @@ class CreateHostSchema(OTPSchema): else: if self.get_errors(): self.specs = { - 'cpu': _cpu, - 'ram': str(parsed_ram), - 'os-ssd': str(parsed_os_ssd), - 'hdd': parsed_hdd + "cpu": _cpu, + "ram": str(parsed_ram), + "os-ssd": str(parsed_os_ssd), + "hdd": parsed_hdd, } def validation(self): if self.realm.value != "ungleich-admin": - self.add_error("Invalid Credentials/Insufficient Permission") + self.add_error( + "Invalid Credentials/Insufficient Permission" + ) # VM Operations @@ -198,13 +232,15 @@ class CreateHostSchema(OTPSchema): class CreateVMSchema(OTPSchema): def __init__(self, data): - self.parsed_specs = {} - # Fields self.specs = Field("specs", dict, data.get("specs", KeyError)) - self.vm_name = Field("vm_name", str, data.get("vm_name", KeyError)) + self.vm_name = Field( + "vm_name", str, data.get("vm_name", KeyError) + ) self.image = Field("image", str, data.get("image", KeyError)) - self.network = Field("network", list, data.get("network", KeyError)) + self.network = Field( + "network", list, data.get("network", KeyError) + ) # Validation self.image.validation = self.image_validation @@ -218,16 +254,25 @@ class CreateVMSchema(OTPSchema): def image_validation(self): try: - image_uuid = helper.resolve_image_name(self.image.value, etcd_client) + image_uuid = helper.resolve_image_name( + self.image.value, shared.etcd_client + ) except Exception as e: + logger.exception( + "Cannot resolve image name = %s", self.image.value + ) self.add_error(str(e)) else: self.image_uuid = image_uuid def vm_name_validation(self): - if resolve_vm_name(name=self.vm_name.value, owner=self.name.value): + if resolve_vm_name( + name=self.vm_name.value, owner=self.name.value + ): self.add_error( - 'VM with same name "{}" already exists'.format(self.vm_name.value) + 'VM with same name "{}" already exists'.format( + self.vm_name.value + ) ) def network_validation(self): @@ -235,34 +280,48 @@ class CreateVMSchema(OTPSchema): if _network: for net in _network: - network = etcd_client.get(os.path.join(env_vars.get('NETWORK_PREFIX'), - self.name.value, - net), value_in_json=True) + network = shared.etcd_client.get( + os.path.join( + shared.shared.shared.shared.shared.settings["etcd"]["network_prefix"], + self.name.value, + net, + ), + value_in_json=True, + ) if not network: - self.add_error("Network with name {} does not exists" \ - .format(net)) + self.add_error( + "Network with name {} does not exists".format( + net + ) + ) def specs_validation(self): ALLOWED_BASE = 10 - _cpu = self.specs.value.get('cpu', KeyError) - _ram = self.specs.value.get('ram', KeyError) - _os_ssd = self.specs.value.get('os-ssd', KeyError) - _hdd = self.specs.value.get('hdd', KeyError) + _cpu = self.specs.value.get("cpu", KeyError) + _ram = self.specs.value.get("ram", KeyError) + _os_ssd = self.specs.value.get("os-ssd", KeyError) + _hdd = self.specs.value.get("hdd", KeyError) if KeyError in [_cpu, _ram, _os_ssd, _hdd]: - self.add_error("You must specify CPU, RAM and OS-SSD in your specs") + self.add_error( + "You must specify CPU, RAM and OS-SSD in your specs" + ) return None try: parsed_ram = bitmath.parse_string_unsafe(_ram) parsed_os_ssd = bitmath.parse_string_unsafe(_os_ssd) if parsed_ram.base != ALLOWED_BASE: - self.add_error("Your specified RAM is not in correct units") + self.add_error( + "Your specified RAM is not in correct units" + ) if parsed_os_ssd.base != ALLOWED_BASE: - self.add_error("Your specified OS-SSD is not in correct units") + self.add_error( + "Your specified OS-SSD is not in correct units" + ) - if _cpu < 1: + if int(_cpu) < 1: self.add_error("CPU must be atleast 1") if parsed_ram < bitmath.GB(1): @@ -275,7 +334,9 @@ class CreateVMSchema(OTPSchema): for hdd in _hdd: _parsed_hdd = bitmath.parse_string_unsafe(hdd) if _parsed_hdd.base != ALLOWED_BASE: - self.add_error("Your specified HDD is not in correct units") + self.add_error( + "Your specified HDD is not in correct units" + ) break else: parsed_hdd.append(str(_parsed_hdd)) @@ -286,21 +347,24 @@ class CreateVMSchema(OTPSchema): else: if self.get_errors(): self.specs = { - 'cpu': _cpu, - 'ram': str(parsed_ram), - 'os-ssd': str(parsed_os_ssd), - 'hdd': parsed_hdd + "cpu": _cpu, + "ram": str(parsed_ram), + "os-ssd": str(parsed_os_ssd), + "hdd": parsed_hdd, } class VMStatusSchema(OTPSchema): def __init__(self, data): data["uuid"] = ( - resolve_vm_name( - name=data.get("vm_name", None), - owner=(data.get("in_support_of", None) or data.get("name", None)), - ) - or KeyError + resolve_vm_name( + name=data.get("vm_name", None), + owner=( + data.get("in_support_of", None) + or data.get("name", None) + ), + ) + or KeyError ) self.uuid = VmUUIDField(data) @@ -309,9 +373,10 @@ class VMStatusSchema(OTPSchema): super().__init__(data, fields) def validation(self): - vm = vm_pool.get(self.uuid.value) + vm = shared.vm_pool.get(self.uuid.value) if not ( - vm.value["owner"] == self.name.value or self.realm.value == "ungleich-admin" + vm.value["owner"] == self.name.value + or self.realm.value == "ungleich-admin" ): self.add_error("Invalid User") @@ -319,11 +384,14 @@ class VMStatusSchema(OTPSchema): class VmActionSchema(OTPSchema): def __init__(self, data): data["uuid"] = ( - resolve_vm_name( - name=data.get("vm_name", None), - owner=(data.get("in_support_of", None) or data.get("name", None)), - ) - or KeyError + resolve_vm_name( + name=data.get("vm_name", None), + owner=( + data.get("in_support_of", None) + or data.get("name", None) + ), + ) + or KeyError ) self.uuid = VmUUIDField(data) self.action = Field("action", str, data.get("action", KeyError)) @@ -338,20 +406,23 @@ class VmActionSchema(OTPSchema): allowed_actions = ["start", "stop", "delete"] if self.action.value not in allowed_actions: self.add_error( - "Invalid Action. Allowed Actions are {}".format(allowed_actions) + "Invalid Action. Allowed Actions are {}".format( + allowed_actions + ) ) def validation(self): - vm = vm_pool.get(self.uuid.value) + vm = shared.vm_pool.get(self.uuid.value) if not ( - vm.value["owner"] == self.name.value or self.realm.value == "ungleich-admin" + vm.value["owner"] == self.name.value + or self.realm.value == "ungleich-admin" ): self.add_error("Invalid User") if ( - self.action.value == "start" - and vm.status == VMStatus.running - and vm.hostname != "" + self.action.value == "start" + and vm.status == VMStatus.running + and vm.hostname != "" ): self.add_error("VM Already Running") @@ -365,15 +436,20 @@ class VmActionSchema(OTPSchema): class VmMigrationSchema(OTPSchema): def __init__(self, data): data["uuid"] = ( - resolve_vm_name( - name=data.get("vm_name", None), - owner=(data.get("in_support_of", None) or data.get("name", None)), - ) - or KeyError + resolve_vm_name( + name=data.get("vm_name", None), + owner=( + data.get("in_support_of", None) + or data.get("name", None) + ), + ) + or KeyError ) self.uuid = VmUUIDField(data) - self.destination = Field("destination", str, data.get("destination", KeyError)) + self.destination = Field( + "destination", str, data.get("destination", KeyError) + ) self.destination.validation = self.destination_validation @@ -382,31 +458,47 @@ class VmMigrationSchema(OTPSchema): def destination_validation(self): hostname = self.destination.value - host = next(filter(lambda h: h.hostname == hostname, host_pool.hosts), None) + host = next( + filter( + lambda h: h.hostname == hostname, shared.host_pool.hosts + ), + None, + ) if not host: - self.add_error("No Such Host ({}) exists".format(self.destination.value)) + self.add_error( + "No Such Host ({}) exists".format( + self.destination.value + ) + ) elif host.status != HostStatus.alive: self.add_error("Destination Host is dead") else: self.destination.value = host.key def validation(self): - vm = vm_pool.get(self.uuid.value) + vm = shared.vm_pool.get(self.uuid.value) if not ( - vm.value["owner"] == self.name.value or self.realm.value == "ungleich-admin" + vm.value["owner"] == self.name.value + or self.realm.value == "ungleich-admin" ): self.add_error("Invalid User") if vm.status != VMStatus.running: self.add_error("Can't migrate non-running VM") - if vm.hostname == os.path.join(env_vars.get('HOST_PREFIX'), self.destination.value): - self.add_error("Destination host couldn't be same as Source Host") + if vm.hostname == os.path.join( + shared.shared.shared.shared.shared.settings["etcd"]["host_prefix"], self.destination.value + ): + self.add_error( + "Destination host couldn't be same as Source Host" + ) class AddSSHSchema(OTPSchema): def __init__(self, data): - self.key_name = Field("key_name", str, data.get("key_name", KeyError)) + self.key_name = Field( + "key_name", str, data.get("key_name", KeyError) + ) self.key = Field("key", str, data.get("key_name", KeyError)) fields = [self.key_name, self.key] @@ -415,7 +507,9 @@ class AddSSHSchema(OTPSchema): class RemoveSSHSchema(OTPSchema): def __init__(self, data): - self.key_name = Field("key_name", str, data.get("key_name", KeyError)) + self.key_name = Field( + "key_name", str, data.get("key_name", KeyError) + ) fields = [self.key_name] super().__init__(data=data, fields=fields) @@ -423,7 +517,9 @@ class RemoveSSHSchema(OTPSchema): class GetSSHSchema(OTPSchema): def __init__(self, data): - self.key_name = Field("key_name", str, data.get("key_name", None)) + self.key_name = Field( + "key_name", str, data.get("key_name", None) + ) fields = [self.key_name] super().__init__(data=data, fields=fields) @@ -442,15 +538,20 @@ class CreateNetwork(OTPSchema): super().__init__(data, fields=fields) def network_name_validation(self): - network = etcd_client.get(os.path.join(env_vars.get('NETWORK_PREFIX'), - self.name.value, - self.network_name.value), - value_in_json=True) + key = os.path.join(shared.shared.shared.shared.shared.settings["etcd"]["network_prefix"], self.name.value, self.network_name.value) + network = shared.etcd_client.get(key, value_in_json=True) if network: - self.add_error("Network with name {} already exists" \ - .format(self.network_name.value)) + self.add_error( + "Network with name {} already exists".format( + self.network_name.value + ) + ) def network_type_validation(self): supported_network_types = ["vxlan"] if self.type.value not in supported_network_types: - self.add_error("Unsupported Network Type. Supported network types are {}".format(supported_network_types)) + self.add_error( + "Unsupported Network Type. Supported network types are {}".format( + supported_network_types + ) + ) diff --git a/ucloud/docs/source/__init__.py b/uncloud/cli/__init__.py similarity index 100% rename from ucloud/docs/source/__init__.py rename to uncloud/cli/__init__.py diff --git a/uncloud/cli/helper.py b/uncloud/cli/helper.py new file mode 100644 index 0000000..51a4355 --- /dev/null +++ b/uncloud/cli/helper.py @@ -0,0 +1,46 @@ +import requests +import json +import argparse +import binascii + +from pyotp import TOTP +from os.path import join as join_path +from uncloud.common.shared import shared + + +def get_otp_parser(): + otp_parser = argparse.ArgumentParser('otp') + otp_parser.add_argument('--name') + otp_parser.add_argument('--realm') + otp_parser.add_argument('--seed', type=get_token, dest='token', metavar='SEED') + + return otp_parser + + +def load_dump_pretty(content): + if isinstance(content, bytes): + content = content.decode('utf-8') + parsed = json.loads(content) + return json.dumps(parsed, indent=4, sort_keys=True) + + +def make_request(*args, data=None, request_method=requests.post): + try: + r = request_method(join_path(shared.settings['client']['api_server'], *args), json=data) + except requests.exceptions.RequestException: + print('Error occurred while connecting to API server.') + else: + try: + print(load_dump_pretty(r.content)) + except Exception: + print('Error occurred while getting output from api server.') + + +def get_token(seed): + if seed is not None: + try: + token = TOTP(seed).now() + except binascii.Error: + raise argparse.ArgumentTypeError('Invalid seed') + else: + return token diff --git a/uncloud/cli/host.py b/uncloud/cli/host.py new file mode 100644 index 0000000..e912567 --- /dev/null +++ b/uncloud/cli/host.py @@ -0,0 +1,45 @@ +import requests + +from uncloud.cli.helper import make_request, get_otp_parser +from uncloud.common.parser import BaseParser + + +class HostParser(BaseParser): + def __init__(self): + super().__init__('host') + + def create(self, **kwargs): + p = self.subparser.add_parser('create', parents=[get_otp_parser()], **kwargs) + p.add_argument('--hostname', required=True) + p.add_argument('--cpu', required=True, type=int) + p.add_argument('--ram', required=True) + p.add_argument('--os-ssd', required=True) + p.add_argument('--hdd', default=list()) + + def list(self, **kwargs): + self.subparser.add_parser('list', **kwargs) + + +parser = HostParser() +arg_parser = parser.arg_parser + + +def main(**kwargs): + subcommand = kwargs.pop('host_subcommand') + if not subcommand: + arg_parser.print_help() + else: + request_method = requests.post + data = None + if subcommand == 'create': + kwargs['specs'] = { + 'cpu': kwargs.pop('cpu'), + 'ram': kwargs.pop('ram'), + 'os-ssd': kwargs.pop('os_ssd'), + 'hdd': kwargs.pop('hdd') + } + data = kwargs + elif subcommand == 'list': + request_method = requests.get + + make_request('host', subcommand, data=data, request_method=request_method) diff --git a/uncloud/cli/image.py b/uncloud/cli/image.py new file mode 100644 index 0000000..2f59c32 --- /dev/null +++ b/uncloud/cli/image.py @@ -0,0 +1,38 @@ +import requests + +from uncloud.cli.helper import make_request +from uncloud.common.parser import BaseParser + + +class ImageParser(BaseParser): + def __init__(self): + super().__init__('image') + + def create(self, **kwargs): + p = self.subparser.add_parser('create', **kwargs) + p.add_argument('--name', required=True) + p.add_argument('--uuid', required=True) + p.add_argument('--image-store', required=True, dest='image_store') + + def list(self, **kwargs): + self.subparser.add_parser('list', **kwargs) + + +parser = ImageParser() +arg_parser = parser.arg_parser + + +def main(**kwargs): + subcommand = kwargs.pop('image_subcommand') + if not subcommand: + arg_parser.print_help() + else: + data = None + request_method = requests.post + if subcommand == 'list': + subcommand = 'list-public' + request_method = requests.get + elif subcommand == 'create': + data = kwargs + + make_request('image', subcommand, data=data, request_method=request_method) diff --git a/uncloud/cli/main.py b/uncloud/cli/main.py new file mode 100644 index 0000000..9a42497 --- /dev/null +++ b/uncloud/cli/main.py @@ -0,0 +1,23 @@ +#!/usr/bin/env python3 + +import argparse +import importlib + +arg_parser = argparse.ArgumentParser('cli', add_help=False) +subparser = arg_parser.add_subparsers(dest='subcommand') + +for component in ['user', 'host', 'image', 'network', 'vm']: + module = importlib.import_module('uncloud.cli.{}'.format(component)) + parser = getattr(module, 'arg_parser') + subparser.add_parser(name=parser.prog, parents=[parser]) + + +def main(arguments): + if not arguments['subcommand']: + arg_parser.print_help() + else: + name = arguments.pop('subcommand') + arguments.pop('debug') + mod = importlib.import_module('uncloud.cli.{}'.format(name)) + _main = getattr(mod, 'main') + _main(**arguments) diff --git a/uncloud/cli/network.py b/uncloud/cli/network.py new file mode 100644 index 0000000..55798bf --- /dev/null +++ b/uncloud/cli/network.py @@ -0,0 +1,32 @@ +import requests + +from uncloud.cli.helper import make_request, get_otp_parser +from uncloud.common.parser import BaseParser + + +class NetworkParser(BaseParser): + def __init__(self): + super().__init__('network') + + def create(self, **kwargs): + p = self.subparser.add_parser('create', parents=[get_otp_parser()], **kwargs) + p.add_argument('--network-name', required=True) + p.add_argument('--network-type', required=True, dest='type') + p.add_argument('--user', action='store_true') + + +parser = NetworkParser() +arg_parser = parser.arg_parser + + +def main(**kwargs): + subcommand = kwargs.pop('network_subcommand') + if not subcommand: + arg_parser.print_help() + else: + data = None + request_method = requests.post + if subcommand == 'create': + data = kwargs + + make_request('network', subcommand, data=data, request_method=request_method) diff --git a/uncloud/cli/user.py b/uncloud/cli/user.py new file mode 100755 index 0000000..3a4cc4e --- /dev/null +++ b/uncloud/cli/user.py @@ -0,0 +1,41 @@ +from uncloud.cli.helper import make_request, get_otp_parser +from uncloud.common.parser import BaseParser + + +class UserParser(BaseParser): + def __init__(self): + super().__init__('user') + + def files(self, **kwargs): + self.subparser.add_parser('files', parents=[get_otp_parser()], **kwargs) + + def vms(self, **kwargs): + self.subparser.add_parser('vms', parents=[get_otp_parser()], **kwargs) + + def networks(self, **kwargs): + self.subparser.add_parser('networks', parents=[get_otp_parser()], **kwargs) + + def add_ssh(self, **kwargs): + p = self.subparser.add_parser('add-ssh', parents=[get_otp_parser()], **kwargs) + p.add_argument('--key-name', required=True) + p.add_argument('--key', required=True) + + def get_ssh(self, **kwargs): + p = self.subparser.add_parser('get-ssh', parents=[get_otp_parser()], **kwargs) + p.add_argument('--key-name', default='') + + def remove_ssh(self, **kwargs): + p = self.subparser.add_parser('remove-ssh', parents=[get_otp_parser()], **kwargs) + p.add_argument('--key-name', required=True) + + +parser = UserParser() +arg_parser = parser.arg_parser + + +def main(**kwargs): + subcommand = kwargs.pop('user_subcommand') + if not subcommand: + arg_parser.print_help() + else: + make_request('user', subcommand, data=kwargs) diff --git a/uncloud/cli/vm.py b/uncloud/cli/vm.py new file mode 100644 index 0000000..396530e --- /dev/null +++ b/uncloud/cli/vm.py @@ -0,0 +1,62 @@ +from uncloud.common.parser import BaseParser +from uncloud.cli.helper import make_request, get_otp_parser + + +class VMParser(BaseParser): + def __init__(self): + super().__init__('vm') + + def start(self, **args): + p = self.subparser.add_parser('start', parents=[get_otp_parser()], **args) + p.add_argument('--vm-name', required=True) + + def stop(self, **args): + p = self.subparser.add_parser('stop', parents=[get_otp_parser()], **args) + p.add_argument('--vm-name', required=True) + + def status(self, **args): + p = self.subparser.add_parser('status', parents=[get_otp_parser()], **args) + p.add_argument('--vm-name', required=True) + + def delete(self, **args): + p = self.subparser.add_parser('delete', parents=[get_otp_parser()], **args) + p.add_argument('--vm-name', required=True) + + def migrate(self, **args): + p = self.subparser.add_parser('migrate', parents=[get_otp_parser()], **args) + p.add_argument('--vm-name', required=True) + p.add_argument('--destination', required=True) + + def create(self, **args): + p = self.subparser.add_parser('create', parents=[get_otp_parser()], **args) + p.add_argument('--cpu', required=True) + p.add_argument('--ram', required=True) + p.add_argument('--os-ssd', required=True) + p.add_argument('--hdd', action='append', default=list()) + p.add_argument('--image', required=True) + p.add_argument('--network', action='append', default=[]) + p.add_argument('--vm-name', required=True) + + +parser = VMParser() +arg_parser = parser.arg_parser + + +def main(**kwargs): + subcommand = kwargs.pop('vm_subcommand') + if not subcommand: + arg_parser.print_help() + else: + data = kwargs + endpoint = subcommand + if subcommand in ['start', 'stop', 'delete']: + endpoint = 'action' + data['action'] = subcommand + elif subcommand == 'create': + kwargs['specs'] = { + 'cpu': kwargs.pop('cpu'), + 'ram': kwargs.pop('ram'), + 'os-ssd': kwargs.pop('os_ssd'), + 'hdd': kwargs.pop('hdd') + } + make_request('vm', endpoint, data=data) diff --git a/ucloud/common/__init__.py b/uncloud/common/__init__.py similarity index 100% rename from ucloud/common/__init__.py rename to uncloud/common/__init__.py diff --git a/ucloud/common/classes.py b/uncloud/common/classes.py similarity index 93% rename from ucloud/common/classes.py rename to uncloud/common/classes.py index 2eae809..29dffd4 100644 --- a/ucloud/common/classes.py +++ b/uncloud/common/classes.py @@ -1,4 +1,4 @@ -from etcd3_wrapper import EtcdEntry +from .etcd_wrapper import EtcdEntry class SpecificEtcdEntryBase: diff --git a/uncloud/common/cli.py b/uncloud/common/cli.py new file mode 100644 index 0000000..3d3c248 --- /dev/null +++ b/uncloud/common/cli.py @@ -0,0 +1,26 @@ +from uncloud.common.shared import shared +from pyotp import TOTP + + +def get_token(seed): + if seed is not None: + try: + token = TOTP(seed).now() + except Exception: + raise Exception('Invalid seed') + else: + return token + + +def resolve_otp_credentials(kwargs): + d = { + 'name': shared.settings['client']['name'], + 'realm': shared.settings['client']['realm'], + 'token': get_token(shared.settings['client']['seed']) + } + + for k, v in d.items(): + if k in kwargs and kwargs[k] is None: + kwargs.update({k: v}) + + return d diff --git a/ucloud/common/counters.py b/uncloud/common/counters.py similarity index 91% rename from ucloud/common/counters.py rename to uncloud/common/counters.py index 066a870..2d4a8e9 100644 --- a/ucloud/common/counters.py +++ b/uncloud/common/counters.py @@ -1,4 +1,4 @@ -from etcd3_wrapper import Etcd3Wrapper +from .etcd_wrapper import Etcd3Wrapper def increment_etcd_counter(etcd_client: Etcd3Wrapper, key): diff --git a/uncloud/common/etcd_wrapper.py b/uncloud/common/etcd_wrapper.py new file mode 100644 index 0000000..38471ab --- /dev/null +++ b/uncloud/common/etcd_wrapper.py @@ -0,0 +1,75 @@ +import etcd3 +import json + +from functools import wraps + +from uncloud import UncloudException +from uncloud.common import logger + + +class EtcdEntry: + def __init__(self, meta_or_key, value, value_in_json=False): + if hasattr(meta_or_key, 'key'): + # if meta has attr 'key' then get it + self.key = meta_or_key.key.decode('utf-8') + else: + # otherwise meta is the 'key' + self.key = meta_or_key + self.value = value.decode('utf-8') + + if value_in_json: + self.value = json.loads(self.value) + + +def readable_errors(func): + @wraps(func) + def wrapper(*args, **kwargs): + try: + return func(*args, **kwargs) + except etcd3.exceptions.ConnectionFailedError: + raise UncloudException('Cannot connect to etcd: is etcd running as configured in uncloud.conf?') + except etcd3.exceptions.ConnectionTimeoutError as err: + raise etcd3.exceptions.ConnectionTimeoutError('etcd connection timeout.') from err + except Exception: + logger.exception('Some etcd error occured. See syslog for details.') + + return wrapper + + +class Etcd3Wrapper: + @readable_errors + def __init__(self, *args, **kwargs): + self.client = etcd3.client(*args, **kwargs) + + @readable_errors + def get(self, *args, value_in_json=False, **kwargs): + _value, _key = self.client.get(*args, **kwargs) + if _key is None or _value is None: + return None + return EtcdEntry(_key, _value, value_in_json=value_in_json) + + @readable_errors + def put(self, *args, value_in_json=False, **kwargs): + _key, _value = args + if value_in_json: + _value = json.dumps(_value) + + if not isinstance(_key, str): + _key = _key.decode('utf-8') + + return self.client.put(_key, _value, **kwargs) + + @readable_errors + def get_prefix(self, *args, value_in_json=False, raise_exception=True, **kwargs): + event_iterator = self.client.get_prefix(*args, **kwargs) + for e in event_iterator: + yield EtcdEntry(*e[::-1], value_in_json=value_in_json) + + @readable_errors + def watch_prefix(self, key, raise_exception=True, value_in_json=False): + event_iterator, cancel = self.client.watch_prefix(key) + for e in event_iterator: + if hasattr(e, '_event'): + e = e._event + if e.type == e.PUT: + yield EtcdEntry(e.kv.key, e.kv.value, value_in_json=value_in_json) diff --git a/ucloud/common/host.py b/uncloud/common/host.py similarity index 85% rename from ucloud/common/host.py rename to uncloud/common/host.py index ccbf7a8..f7bb7d5 100644 --- a/ucloud/common/host.py +++ b/uncloud/common/host.py @@ -7,7 +7,7 @@ from .classes import SpecificEtcdEntryBase class HostStatus: - """Possible Statuses of ucloud host.""" + """Possible Statuses of uncloud host.""" alive = "ALIVE" dead = "DEAD" @@ -26,11 +26,13 @@ class HostEntry(SpecificEtcdEntryBase): def update_heartbeat(self): self.status = HostStatus.alive - self.last_heartbeat = time.strftime("%Y-%m-%d %H:%M:%S") + self.last_heartbeat = datetime.utcnow().strftime("%Y-%m-%d %H:%M:%S") def is_alive(self): - last_heartbeat = datetime.strptime(self.last_heartbeat, "%Y-%m-%d %H:%M:%S") - delta = datetime.now() - last_heartbeat + last_heartbeat = datetime.strptime( + self.last_heartbeat, "%Y-%m-%d %H:%M:%S" + ) + delta = datetime.utcnow() - last_heartbeat if delta.total_seconds() > 60: return False return True diff --git a/uncloud/common/network.py b/uncloud/common/network.py new file mode 100644 index 0000000..32f6951 --- /dev/null +++ b/uncloud/common/network.py @@ -0,0 +1,70 @@ +import subprocess as sp +import random +import logging + +logger = logging.getLogger(__name__) + + +def random_bytes(num=6): + return [random.randrange(256) for _ in range(num)] + + +def generate_mac( + uaa=False, multicast=False, oui=None, separator=":", byte_fmt="%02x" +): + mac = random_bytes() + if oui: + if type(oui) == str: + oui = [int(chunk) for chunk in oui.split(separator)] + mac = oui + random_bytes(num=6 - len(oui)) + else: + if multicast: + mac[0] |= 1 # set bit 0 + else: + mac[0] &= ~1 # clear bit 0 + if uaa: + mac[0] &= ~(1 << 1) # clear bit 1 + else: + mac[0] |= 1 << 1 # set bit 1 + return separator.join(byte_fmt % b for b in mac) + + +def create_dev(script, _id, dev, ip=None): + command = [ + "sudo", + "-p", + "Enter password to create network devices for vm: ", + script, + str(_id), + dev, + ] + if ip: + command.append(ip) + try: + output = sp.check_output(command, stderr=sp.PIPE) + except Exception: + logger.exception("Creation of interface %s failed.", dev) + return None + else: + return output.decode("utf-8").strip() + + +def delete_network_interface(iface): + try: + sp.check_output( + [ + "sudo", + "-p", + "Enter password to remove {} network device: ".format( + iface + ), + "ip", + "link", + "del", + iface, + ], + stderr=sp.PIPE, + ) + except Exception: + logger.exception("Interface %s Deletion failed", iface) + diff --git a/uncloud/common/parser.py b/uncloud/common/parser.py new file mode 100644 index 0000000..576f0e7 --- /dev/null +++ b/uncloud/common/parser.py @@ -0,0 +1,13 @@ +import argparse + + +class BaseParser: + def __init__(self, command): + self.arg_parser = argparse.ArgumentParser(command, add_help=False) + self.subparser = self.arg_parser.add_subparsers(dest='{}_subcommand'.format(command)) + self.common_args = {'add_help': False} + + methods = [attr for attr in dir(self) if not attr.startswith('__') + and type(getattr(self, attr)).__name__ == 'method'] + for method in methods: + getattr(self, method)(**self.common_args) diff --git a/ucloud/common/request.py b/uncloud/common/request.py similarity index 74% rename from ucloud/common/request.py rename to uncloud/common/request.py index cadac80..cb0add5 100644 --- a/ucloud/common/request.py +++ b/uncloud/common/request.py @@ -2,9 +2,8 @@ import json from os.path import join from uuid import uuid4 -from etcd3_wrapper.etcd3_wrapper import PsuedoEtcdEntry - -from .classes import SpecificEtcdEntryBase +from uncloud.common.etcd_wrapper import EtcdEntry +from uncloud.common.classes import SpecificEtcdEntryBase class RequestType: @@ -18,8 +17,9 @@ class RequestType: class RequestEntry(SpecificEtcdEntryBase): - def __init__(self, e): + self.destination_sock_path = None + self.destination_host_key = None self.type = None # type: str self.migration = None # type: bool self.destination = None # type: str @@ -29,8 +29,8 @@ class RequestEntry(SpecificEtcdEntryBase): @classmethod def from_scratch(cls, request_prefix, **kwargs): - e = PsuedoEtcdEntry(join(request_prefix, uuid4().hex), - value=json.dumps(kwargs).encode("utf-8"), value_in_json=True) + e = EtcdEntry(meta_or_key=join(request_prefix, uuid4().hex), + value=json.dumps(kwargs).encode('utf-8'), value_in_json=True) return cls(e) diff --git a/uncloud/common/schemas.py b/uncloud/common/schemas.py new file mode 100644 index 0000000..04978a5 --- /dev/null +++ b/uncloud/common/schemas.py @@ -0,0 +1,41 @@ +import bitmath + +from marshmallow import fields, Schema + + +class StorageUnit(fields.Field): + def _serialize(self, value, attr, obj, **kwargs): + return str(value) + + def _deserialize(self, value, attr, data, **kwargs): + return bitmath.parse_string_unsafe(value) + + +class SpecsSchema(Schema): + cpu = fields.Int() + ram = StorageUnit() + os_ssd = StorageUnit(data_key="os-ssd", attribute="os-ssd") + hdd = fields.List(StorageUnit()) + + +class VMSchema(Schema): + name = fields.Str() + owner = fields.Str() + owner_realm = fields.Str() + specs = fields.Nested(SpecsSchema) + status = fields.Str() + log = fields.List(fields.Str()) + vnc_socket = fields.Str() + image_uuid = fields.Str() + hostname = fields.Str() + metadata = fields.Dict() + network = fields.List( + fields.Tuple((fields.Str(), fields.Str(), fields.Int())) + ) + in_migration = fields.Bool() + + +class NetworkSchema(Schema): + _id = fields.Int(data_key="id", attribute="id") + _type = fields.Str(data_key="type", attribute="type") + ipv6 = fields.Str() diff --git a/uncloud/common/settings.py b/uncloud/common/settings.py new file mode 100644 index 0000000..8503f42 --- /dev/null +++ b/uncloud/common/settings.py @@ -0,0 +1,136 @@ +import configparser +import logging +import sys +import os + +from datetime import datetime +from uncloud.common.etcd_wrapper import Etcd3Wrapper +from os.path import join as join_path + +logger = logging.getLogger(__name__) +settings = None + + +class CustomConfigParser(configparser.RawConfigParser): + def __getitem__(self, key): + try: + result = super().__getitem__(key) + except KeyError as err: + raise KeyError( + 'Key \'{}\' not found in configuration. Make sure you configure uncloud.'.format( + key + ) + ) from err + else: + return result + + +class Settings(object): + def __init__(self, conf_dir, seed_value=None): + conf_name = 'uncloud.conf' + self.config_file = join_path(conf_dir, conf_name) + + # this is used to cache config from etcd for 1 minutes. Without this we + # would make a lot of requests to etcd which slows down everything. + self.last_config_update = datetime.fromtimestamp(0) + + self.config_parser = CustomConfigParser(allow_no_value=True) + self.config_parser.add_section('etcd') + self.config_parser.set('etcd', 'base_prefix', '/') + + if os.access(self.config_file, os.R_OK): + self.config_parser.read(self.config_file) + else: + raise FileNotFoundError('Config file %s not found!', self.config_file) + self.config_key = join_path(self['etcd']['base_prefix'] + 'uncloud/config/') + + self.read_internal_values() + + if seed_value is None: + seed_value = dict() + + self.config_parser.read_dict(seed_value) + + def get_etcd_client(self): + args = tuple() + try: + kwargs = { + 'host': self.config_parser.get('etcd', 'url'), + 'port': self.config_parser.get('etcd', 'port'), + 'ca_cert': self.config_parser.get('etcd', 'ca_cert'), + 'cert_cert': self.config_parser.get('etcd', 'cert_cert'), + 'cert_key': self.config_parser.get('etcd', 'cert_key'), + } + except configparser.Error as err: + raise configparser.Error( + '{} in config file {}'.format( + err.message, self.config_file + ) + ) from err + else: + try: + wrapper = Etcd3Wrapper(*args, **kwargs) + except Exception as err: + logger.error( + 'etcd connection not successfull. Please check your config file.' + '\nDetails: %s\netcd connection parameters: %s', + err, + kwargs, + ) + sys.exit(1) + else: + return wrapper + + def read_internal_values(self): + base_prefix = self['etcd']['base_prefix'] + self.config_parser.read_dict( + { + 'etcd': { + 'file_prefix': join_path(base_prefix, 'files/'), + 'host_prefix': join_path(base_prefix, 'hosts/'), + 'image_prefix': join_path(base_prefix, 'images/'), + 'image_store_prefix': join_path(base_prefix, 'imagestore/'), + 'network_prefix': join_path(base_prefix, 'networks/'), + 'request_prefix': join_path(base_prefix, 'requests/'), + 'user_prefix': join_path(base_prefix, 'users/'), + 'vm_prefix': join_path(base_prefix, 'vms/'), + 'vxlan_counter': join_path(base_prefix, 'counters/vxlan'), + 'tap_counter': join_path(base_prefix, 'counters/tap') + } + } + ) + + def read_config_file_values(self, config_file): + try: + # Trying to read configuration file + with open(config_file) as config_file_handle: + self.config_parser.read_file(config_file_handle) + except FileNotFoundError: + sys.exit('Configuration file {} not found!'.format(config_file)) + except Exception as err: + logger.exception(err) + sys.exit('Error occurred while reading configuration file') + + def read_values_from_etcd(self): + etcd_client = self.get_etcd_client() + if (datetime.utcnow() - self.last_config_update).total_seconds() > 60: + config_from_etcd = etcd_client.get(self.config_key, value_in_json=True) + if config_from_etcd: + self.config_parser.read_dict(config_from_etcd.value) + self.last_config_update = datetime.utcnow() + else: + raise KeyError('Key \'{}\' not found in etcd. Please configure uncloud.'.format(self.config_key)) + + def __getitem__(self, key): + # Allow failing to read from etcd if we have + # it locally + if key not in self.config_parser.sections(): + try: + self.read_values_from_etcd() + except KeyError: + pass + return self.config_parser[key] + + +def get_settings(): + return settings diff --git a/uncloud/common/shared.py b/uncloud/common/shared.py new file mode 100644 index 0000000..aea7cbc --- /dev/null +++ b/uncloud/common/shared.py @@ -0,0 +1,34 @@ +from uncloud.common.settings import get_settings +from uncloud.common.vm import VmPool +from uncloud.common.host import HostPool +from uncloud.common.request import RequestPool +import uncloud.common.storage_handlers as storage_handlers + + +class Shared: + @property + def settings(self): + return get_settings() + + @property + def etcd_client(self): + return self.settings.get_etcd_client() + + @property + def host_pool(self): + return HostPool(self.etcd_client, self.settings["etcd"]["host_prefix"]) + + @property + def vm_pool(self): + return VmPool(self.etcd_client, self.settings["etcd"]["vm_prefix"]) + + @property + def request_pool(self): + return RequestPool(self.etcd_client, self.settings["etcd"]["request_prefix"]) + + @property + def storage_handler(self): + return storage_handlers.get_storage_handler() + + +shared = Shared() diff --git a/ucloud/common/storage_handlers.py b/uncloud/common/storage_handlers.py similarity index 63% rename from ucloud/common/storage_handlers.py rename to uncloud/common/storage_handlers.py index 8b1097a..58c2dc2 100644 --- a/ucloud/common/storage_handlers.py +++ b/uncloud/common/storage_handlers.py @@ -6,17 +6,20 @@ import stat from abc import ABC from . import logger from os.path import join as join_path +import uncloud.common.shared as shared class ImageStorageHandler(ABC): + handler_name = "base" + def __init__(self, image_base, vm_base): self.image_base = image_base self.vm_base = vm_base def import_image(self, image_src, image_dest, protect=False): """Put an image at the destination - :param src: An Image file - :param dest: A path where :param src: is to be put. + :param image_src: An Image file + :param image_dest: A path where :param src: is to be put. :param protect: If protect is true then the dest is protect (readonly etc) The obj must exist on filesystem. """ @@ -26,8 +29,8 @@ class ImageStorageHandler(ABC): def make_vm_image(self, image_path, path): """Copy image from src to dest - :param src: A path - :param dest: A path + :param image_path: A path + :param path: A path src and destination must be on same storage system i.e both on file system or both on CEPH etc. """ @@ -43,14 +46,17 @@ class ImageStorageHandler(ABC): def delete_vm_image(self, path): raise NotImplementedError() - def execute_command(self, command, report=True): + def execute_command(self, command, report=True, error_origin=None): + if not error_origin: + error_origin = self.handler_name + command = list(map(str, command)) try: - output = sp.check_output(command, stderr=sp.PIPE) - except Exception as e: + sp.check_output(command, stderr=sp.PIPE) + except sp.CalledProcessError as e: + _stderr = e.stderr.decode("utf-8").strip() if report: - print(e) - logger.exception(e) + logger.exception("%s:- %s", error_origin, _stderr) return False return True @@ -65,12 +71,16 @@ class ImageStorageHandler(ABC): class FileSystemBasedImageStorageHandler(ImageStorageHandler): + handler_name = "Filesystem" + def import_image(self, src, dest, protect=False): dest = join_path(self.image_base, dest) try: shutil.copy(src, dest) if protect: - os.chmod(dest, stat.S_IRUSR | stat.S_IRGRP | stat.S_IROTH) + os.chmod( + dest, stat.S_IRUSR | stat.S_IRGRP | stat.S_IROTH + ) except Exception as e: logger.exception(e) return False @@ -80,7 +90,7 @@ class FileSystemBasedImageStorageHandler(ImageStorageHandler): src = join_path(self.image_base, src) dest = join_path(self.vm_base, dest) try: - shutil.copy(src, dest) + shutil.copyfile(src, dest) except Exception as e: logger.exception(e) return False @@ -88,7 +98,14 @@ class FileSystemBasedImageStorageHandler(ImageStorageHandler): def resize_vm_image(self, path, size): path = join_path(self.vm_base, path) - command = ["qemu-img", "resize", "-f", "raw", path, "{}M".format(size)] + command = [ + "qemu-img", + "resize", + "-f", + "raw", + path, + "{}M".format(size), + ] if self.execute_command(command): return True else: @@ -117,17 +134,33 @@ class FileSystemBasedImageStorageHandler(ImageStorageHandler): class CEPHBasedImageStorageHandler(ImageStorageHandler): + handler_name = "Ceph" + def import_image(self, src, dest, protect=False): dest = join_path(self.image_base, dest) - command = ["rbd", "import", src, dest] + import_command = ["rbd", "import", src, dest] + commands = [import_command] if protect: - snap_create_command = ["rbd", "snap", "create", "{}@protected".format(dest)] - snap_protect_command = ["rbd", "snap", "protect", "{}@protected".format(dest)] + snap_create_command = [ + "rbd", + "snap", + "create", + "{}@protected".format(dest), + ] + snap_protect_command = [ + "rbd", + "snap", + "protect", + "{}@protected".format(dest), + ] + commands.append(snap_create_command) + commands.append(snap_protect_command) - return self.execute_command(command) and self.execute_command(snap_create_command) and\ - self.execute_command(snap_protect_command) + result = True + for command in commands: + result = result and self.execute_command(command) - return self.execute_command(command) + return result def make_vm_image(self, src, dest): src = join_path(self.image_base, src) @@ -156,3 +189,19 @@ class CEPHBasedImageStorageHandler(ImageStorageHandler): path = join_path(self.vm_base, path) command = ["rbd", "info", path] return self.execute_command(command, report=False) + + +def get_storage_handler(): + __storage_backend = shared.shared.settings["storage"]["storage_backend"] + if __storage_backend == "filesystem": + return FileSystemBasedImageStorageHandler( + vm_base=shared.shared.settings["storage"]["vm_dir"], + image_base=shared.shared.settings["storage"]["image_dir"], + ) + elif __storage_backend == "ceph": + return CEPHBasedImageStorageHandler( + vm_base=shared.shared.settings["storage"]["ceph_vm_pool"], + image_base=shared.shared.settings["storage"]["ceph_image_pool"], + ) + else: + raise Exception("Unknown Image Storage Handler") \ No newline at end of file diff --git a/ucloud/common/vm.py b/uncloud/common/vm.py similarity index 92% rename from ucloud/common/vm.py rename to uncloud/common/vm.py index 0fb5cea..d11046d 100644 --- a/ucloud/common/vm.py +++ b/uncloud/common/vm.py @@ -12,8 +12,13 @@ class VMStatus: error = "ERROR" # An error occurred that cannot be resolved automatically -class VMEntry(SpecificEtcdEntryBase): +def declare_stopped(vm): + vm["hostname"] = "" + vm["in_migration"] = False + vm["status"] = VMStatus.stopped + +class VMEntry(SpecificEtcdEntryBase): def __init__(self, e): self.owner = None # type: str self.specs = None # type: dict @@ -42,7 +47,9 @@ class VMEntry(SpecificEtcdEntryBase): def add_log(self, msg): self.log = self.log[:5] - self.log.append("{} - {}".format(datetime.now().isoformat(), msg)) + self.log.append( + "{} - {}".format(datetime.now().isoformat(), msg) + ) class VmPool: diff --git a/ucloud/metadata/__init__.py b/uncloud/configure/__init__.py similarity index 100% rename from ucloud/metadata/__init__.py rename to uncloud/configure/__init__.py diff --git a/uncloud/configure/main.py b/uncloud/configure/main.py new file mode 100644 index 0000000..87f5752 --- /dev/null +++ b/uncloud/configure/main.py @@ -0,0 +1,57 @@ +import os +import argparse + +from uncloud.common.shared import shared + +arg_parser = argparse.ArgumentParser('configure', add_help=False) +configure_subparsers = arg_parser.add_subparsers(dest='subcommand') + +otp_parser = configure_subparsers.add_parser('otp') +otp_parser.add_argument('--verification-controller-url', required=True, metavar='URL') +otp_parser.add_argument('--auth-name', required=True, metavar='OTP-NAME') +otp_parser.add_argument('--auth-realm', required=True, metavar='OTP-REALM') +otp_parser.add_argument('--auth-seed', required=True, metavar='OTP-SEED') + +network_parser = configure_subparsers.add_parser('network') +network_parser.add_argument('--prefix-length', required=True, type=int) +network_parser.add_argument('--prefix', required=True) +network_parser.add_argument('--vxlan-phy-dev', required=True) + +netbox_parser = configure_subparsers.add_parser('netbox') +netbox_parser.add_argument('--url', required=True) +netbox_parser.add_argument('--token', required=True) + +ssh_parser = configure_subparsers.add_parser('ssh') +ssh_parser.add_argument('--username', default='root') +ssh_parser.add_argument('--private-key-path', default=os.path.expanduser('~/.ssh/id_rsa'),) + +storage_parser = configure_subparsers.add_parser('storage') +storage_parser.add_argument('--file-dir', required=True) +storage_parser_subparsers = storage_parser.add_subparsers(dest='storage_backend') + +filesystem_storage_parser = storage_parser_subparsers.add_parser('filesystem') +filesystem_storage_parser.add_argument('--vm-dir', required=True) +filesystem_storage_parser.add_argument('--image-dir', required=True) + +ceph_storage_parser = storage_parser_subparsers.add_parser('ceph') +ceph_storage_parser.add_argument('--ceph-vm-pool', required=True) +ceph_storage_parser.add_argument('--ceph-image-pool', required=True) + + +def update_config(section, kwargs): + uncloud_config = shared.etcd_client.get(shared.settings.config_key, value_in_json=True) + if not uncloud_config: + uncloud_config = {} + else: + uncloud_config = uncloud_config.value + + uncloud_config[section] = kwargs + shared.etcd_client.put(shared.settings.config_key, uncloud_config, value_in_json=True) + + +def main(arguments): + subcommand = arguments['subcommand'] + if not subcommand: + arg_parser.print_help() + else: + update_config(subcommand, arguments) diff --git a/ucloud/filescanner/__init__.py b/uncloud/filescanner/__init__.py similarity index 100% rename from ucloud/filescanner/__init__.py rename to uncloud/filescanner/__init__.py diff --git a/uncloud/filescanner/main.py b/uncloud/filescanner/main.py new file mode 100755 index 0000000..046f915 --- /dev/null +++ b/uncloud/filescanner/main.py @@ -0,0 +1,85 @@ +import glob +import os +import pathlib +import subprocess as sp +import time +import argparse +import bitmath + +from uuid import uuid4 + +from . import logger +from uncloud.common.shared import shared + +arg_parser = argparse.ArgumentParser('filescanner', add_help=False) +arg_parser.add_argument('--hostname', required=True) + + +def sha512sum(file: str): + """Use sha512sum utility to compute sha512 sum of arg:file + + IF arg:file does not exists: + raise FileNotFoundError exception + ELSE IF sum successfully computer: + return computed sha512 sum + ELSE: + return None + """ + if not isinstance(file, str): + raise TypeError + try: + output = sp.check_output(['sha512sum', file], stderr=sp.PIPE) + except sp.CalledProcessError as e: + error = e.stderr.decode('utf-8') + if 'No such file or directory' in error: + raise FileNotFoundError from None + else: + output = output.decode('utf-8').strip() + output = output.split(' ') + return output[0] + return None + + +def track_file(file, base_dir, host): + file_path = file.relative_to(base_dir) + file_str = str(file) + # Get Username + try: + owner = file_path.parts[0] + except IndexError: + pass + else: + file_path = file_path.relative_to(owner) + creation_date = time.ctime(os.stat(file_str).st_ctime) + + entry_key = os.path.join(shared.settings['etcd']['file_prefix'], str(uuid4())) + entry_value = { + 'filename': str(file_path), + 'owner': owner, + 'sha512sum': sha512sum(file_str), + 'creation_date': creation_date, + 'size': str(bitmath.Byte(os.path.getsize(file_str)).to_MB()), + 'host': host + } + + logger.info('Tracking %s', file_str) + + shared.etcd_client.put(entry_key, entry_value, value_in_json=True) + + +def main(arguments): + hostname = arguments['hostname'] + base_dir = shared.settings['storage']['file_dir'] + # Recursively Get All Files and Folder below BASE_DIR + files = glob.glob('{}/**'.format(base_dir), recursive=True) + files = [pathlib.Path(f) for f in files if pathlib.Path(f).is_file()] + + # Files that are already tracked + tracked_files = [ + pathlib.Path(os.path.join(base_dir, f.value['owner'], f.value['filename'])) + for f in shared.etcd_client.get_prefix(shared.settings['etcd']['file_prefix'], value_in_json=True) + if f.value['host'] == hostname + ] + untracked_files = set(files) - set(tracked_files) + for file in untracked_files: + track_file(file, base_dir, hostname) diff --git a/ucloud/hack/README.org b/uncloud/hack/README.org similarity index 100% rename from ucloud/hack/README.org rename to uncloud/hack/README.org diff --git a/ucloud/network/__init__.py b/uncloud/hack/__init__.py similarity index 100% rename from ucloud/network/__init__.py rename to uncloud/hack/__init__.py diff --git a/ucloud/hack/conf.d/ucloud-host b/uncloud/hack/conf.d/ucloud-host similarity index 100% rename from ucloud/hack/conf.d/ucloud-host rename to uncloud/hack/conf.d/ucloud-host diff --git a/uncloud/hack/config.py b/uncloud/hack/config.py new file mode 100644 index 0000000..7e2655d --- /dev/null +++ b/uncloud/hack/config.py @@ -0,0 +1,39 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +# +# 2020 Nico Schottelius (nico.schottelius at ungleich.ch) +# +# This file is part of uncloud. +# +# uncloud is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# uncloud is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with uncloud. If not, see . +# +# + +class Config(object): + def __init__(self, arguments): + """ read arguments dicts as a base """ + + self.arguments = arguments + + # Split them so *etcd_args can be used and we can + # iterate over etcd_hosts + self.etcd_hosts = [ arguments['etcd_host'] ] + self.etcd_args = { + 'ca_cert': arguments['etcd_ca_cert'], + 'cert_cert': arguments['etcd_cert_cert'], + 'cert_key': arguments['etcd_cert_key'], +# 'user': None, +# 'password': None + } + self.etcd_prefix = '/nicohack/' diff --git a/uncloud/hack/db.py b/uncloud/hack/db.py new file mode 100644 index 0000000..cb5e490 --- /dev/null +++ b/uncloud/hack/db.py @@ -0,0 +1,113 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +# +# 2020 Nico Schottelius (nico.schottelius at ungleich.ch) +# +# This file is part of uncloud. +# +# uncloud is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# uncloud is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with uncloud. If not, see . +# +# + +import etcd3 +import json +import logging + +from functools import wraps +from uncloud import UncloudException + +log = logging.getLogger(__name__) + + +def readable_errors(func): + @wraps(func) + def wrapper(*args, **kwargs): + try: + return func(*args, **kwargs) + except etcd3.exceptions.ConnectionFailedError as e: + raise UncloudException('Cannot connect to etcd: is etcd running and reachable? {}'.format(e)) + except etcd3.exceptions.ConnectionTimeoutError as e: + raise UncloudException('etcd connection timeout. {}'.format(e)) + + return wrapper + + +class DB(object): + def __init__(self, config, prefix="/"): + self.config = config + + # Root for everything + self.base_prefix= '/nicohack' + + # Can be set from outside + self.prefix = prefix + + self.connect() + + @readable_errors + def connect(self): + self._db_clients = [] + for endpoint in self.config.etcd_hosts: + client = etcd3.client(host=endpoint, **self.config.etcd_args) + self._db_clients.append(client) + + def realkey(self, key): + return "{}{}/{}".format(self.base_prefix, + self.prefix, + key) + + @readable_errors + def get(self, key, as_json=False, **kwargs): + value, _ = self._db_clients[0].get(self.realkey(key), **kwargs) + + if as_json: + value = json.loads(value) + + return value + + + @readable_errors + def set(self, key, value, as_json=False, **kwargs): + if as_json: + value = json.dumps(value) + + # FIXME: iterate over clients in case of failure ? + return self._db_clients[0].put(self.realkey(key), value, **kwargs) + + @readable_errors + def increment(self, key, **kwargs): + print(self.realkey(key)) + + + print("prelock") + lock = self._db_clients[0].lock('/nicohack/foo') + print("prelockacq") + lock.acquire() + print("prelockrelease") + lock.release() + + with self._db_clients[0].lock("/nicohack/mac/last_used_index") as lock: + print("in lock") + pass + +# with self._db_clients[0].lock(self.realkey(key)) as lock:# value = int(self.get(self.realkey(key), **kwargs)) +# self.set(self.realkey(key), str(value + 1), **kwargs) + + +if __name__ == '__main__': + endpoints = [ "https://etcd1.ungleich.ch:2379", + "https://etcd2.ungleich.ch:2379", + "https://etcd3.ungleich.ch:2379" ] + + db = DB(url=endpoints) diff --git a/uncloud/hack/hackcloud/.gitignore b/uncloud/hack/hackcloud/.gitignore new file mode 100644 index 0000000..0ad647b --- /dev/null +++ b/uncloud/hack/hackcloud/.gitignore @@ -0,0 +1,3 @@ +*.iso +radvdpid +foo diff --git a/uncloud/hack/hackcloud/__init__.py b/uncloud/hack/hackcloud/__init__.py new file mode 100644 index 0000000..8b13789 --- /dev/null +++ b/uncloud/hack/hackcloud/__init__.py @@ -0,0 +1 @@ + diff --git a/uncloud/hack/hackcloud/etcd-client.sh b/uncloud/hack/hackcloud/etcd-client.sh new file mode 100644 index 0000000..ab102a5 --- /dev/null +++ b/uncloud/hack/hackcloud/etcd-client.sh @@ -0,0 +1,6 @@ +#!/bin/sh + +etcdctl --cert=$HOME/vcs/ungleich-dot-cdist/files/etcd/nico.pem \ + --key=/home/nico/vcs/ungleich-dot-cdist/files/etcd/nico-key.pem \ + --cacert=$HOME/vcs/ungleich-dot-cdist/files/etcd/ca.pem \ + --endpoints https://etcd1.ungleich.ch:2379,https://etcd2.ungleich.ch:2379,https://etcd3.ungleich.ch:2379 "$@" diff --git a/uncloud/hack/hackcloud/ifdown.sh b/uncloud/hack/hackcloud/ifdown.sh new file mode 100755 index 0000000..5753099 --- /dev/null +++ b/uncloud/hack/hackcloud/ifdown.sh @@ -0,0 +1,3 @@ +#!/bin/sh + +echo $@ diff --git a/uncloud/hack/hackcloud/ifup.sh b/uncloud/hack/hackcloud/ifup.sh new file mode 100755 index 0000000..e0a3ca0 --- /dev/null +++ b/uncloud/hack/hackcloud/ifup.sh @@ -0,0 +1,7 @@ +#!/bin/sh + +dev=$1; shift + +# bridge is setup from outside +ip link set dev "$dev" master ${bridge} +ip link set dev "$dev" up diff --git a/uncloud/hack/hackcloud/mac-last b/uncloud/hack/hackcloud/mac-last new file mode 100644 index 0000000..8c5f254 --- /dev/null +++ b/uncloud/hack/hackcloud/mac-last @@ -0,0 +1 @@ +000000000252 diff --git a/uncloud/hack/hackcloud/mac-prefix b/uncloud/hack/hackcloud/mac-prefix new file mode 100644 index 0000000..5084a2f --- /dev/null +++ b/uncloud/hack/hackcloud/mac-prefix @@ -0,0 +1 @@ +02:00 diff --git a/uncloud/hack/hackcloud/net.sh b/uncloud/hack/hackcloud/net.sh new file mode 100755 index 0000000..4e2bfa1 --- /dev/null +++ b/uncloud/hack/hackcloud/net.sh @@ -0,0 +1,29 @@ +#!/bin/sh + +set -x + +netid=100 +dev=wlp2s0 +dev=wlp0s20f3 +#dev=wlan0 + +ip=2a0a:e5c1:111:888::48/64 +vxlandev=vxlan${netid} +bridgedev=br${netid} + +ip -6 link add ${vxlandev} type vxlan \ + id ${netid} \ + dstport 4789 \ + group ff05::${netid} \ + dev ${dev} \ + ttl 5 + +ip link set ${vxlandev} up + + +ip link add ${bridgedev} type bridge +ip link set ${bridgedev} up + +ip link set ${vxlandev} master ${bridgedev} up + +ip addr add ${ip} dev ${bridgedev} diff --git a/uncloud/hack/hackcloud/nftrules b/uncloud/hack/hackcloud/nftrules new file mode 100644 index 0000000..636c63d --- /dev/null +++ b/uncloud/hack/hackcloud/nftrules @@ -0,0 +1,31 @@ +flush ruleset + +table bridge filter { + chain prerouting { + type filter hook prerouting priority 0; + policy accept; + + ibrname br100 jump br100 + } + + chain br100 { + # Allow all incoming traffic from outside + iifname vxlan100 accept + + # Default blocks: router advertisements, dhcpv6, dhcpv4 + icmpv6 type nd-router-advert drop + ip6 version 6 udp sport 547 drop + ip version 4 udp sport 67 drop + + jump br100_vmlist + drop + } + chain br100_vmlist { + # VM1 + iifname tap1 ether saddr 02:00:f0:a9:c4:4e ip6 saddr 2a0a:e5c1:111:888:0:f0ff:fea9:c44e accept + + # VM2 + iifname v343a-0 ether saddr 02:00:f0:a9:c4:4f ip6 saddr 2a0a:e5c1:111:888:0:f0ff:fea9:c44f accept + iifname v343a-0 ether saddr 02:00:f0:a9:c4:4f ip6 saddr 2a0a:e5c1:111:1234::/64 accept + } +} diff --git a/uncloud/hack/hackcloud/radvd.conf b/uncloud/hack/hackcloud/radvd.conf new file mode 100644 index 0000000..3d8ce4d --- /dev/null +++ b/uncloud/hack/hackcloud/radvd.conf @@ -0,0 +1,13 @@ +interface br100 +{ + AdvSendAdvert on; + MinRtrAdvInterval 3; + MaxRtrAdvInterval 5; + AdvDefaultLifetime 3600; + + prefix 2a0a:e5c1:111:888::/64 { + }; + + RDNSS 2a0a:e5c0::3 2a0a:e5c0::4 { AdvRDNSSLifetime 6000; }; + DNSSL place7.ungleich.ch { AdvDNSSLLifetime 6000; } ; +}; diff --git a/uncloud/hack/hackcloud/radvd.sh b/uncloud/hack/hackcloud/radvd.sh new file mode 100644 index 0000000..9d0e7d1 --- /dev/null +++ b/uncloud/hack/hackcloud/radvd.sh @@ -0,0 +1,3 @@ +#!/bin/sh + +radvd -C ./radvd.conf -n -p ./radvdpid diff --git a/uncloud/hack/hackcloud/vm.sh b/uncloud/hack/hackcloud/vm.sh new file mode 100755 index 0000000..dd9be84 --- /dev/null +++ b/uncloud/hack/hackcloud/vm.sh @@ -0,0 +1,29 @@ +#!/bin/sh + +# if [ $# -ne 1 ]; then +# echo "$0: owner" +# exit 1 +# fi + +qemu=/usr/bin/qemu-system-x86_64 + +accel=kvm +#accel=tcg + +memory=1024 +cores=2 +uuid=$(uuidgen) +mac=$(./mac-gen.py) +owner=nico + +export bridge=br100 + +set -x +$qemu -name "uncloud-${uuid}" \ + -machine pc,accel=${accel} \ + -m ${memory} \ + -smp ${cores} \ + -uuid ${uuid} \ + -drive file=alpine-virt-3.11.2-x86_64.iso,media=cdrom \ + -netdev tap,id=netmain,script=./ifup.sh,downscript=./ifdown.sh \ + -device virtio-net-pci,netdev=netmain,id=net0,mac=${mac} diff --git a/uncloud/hack/mac.py b/uncloud/hack/mac.py new file mode 100755 index 0000000..66286dd --- /dev/null +++ b/uncloud/hack/mac.py @@ -0,0 +1,102 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +# +# 2012 Nico Schottelius (nico-cinv at schottelius.org) +# +# This file is part of cinv. +# +# cinv is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# cinv is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with cinv. If not, see . +# +# + +import argparse +import logging +import os.path +import os +import re +import json + +from uncloud import UncloudException +from uncloud.hack.db import DB + +log = logging.getLogger(__name__) + + +class MAC(object): + def __init__(self, config): + self.config = config + self.no_db = self.config.arguments['no_db'] + if not self.no_db: + self.db = DB(config, prefix="/mac") + + self.prefix = 0x420000000000 + self._number = 0 # Not set by default + + @staticmethod + def validate_mac(mac): + if not re.match(r'([0-9A-F]{2}[-:]){5}[0-9A-F]{2}$', mac, re.I): + raise Error("Not a valid mac address: %s" % mac) + + def last_used_index(self): + if not self.no_db: + value = self.db.get("last_used_index") + if not value: + self.db.set("last_used_index", "0") + value = self.db.get("last_used_index") + + else: + value = "0" + + return int(value) + + def last_used_mac(self): + return self.int_to_mac(self.prefix + self.last_used_index()) + + def to_colon_format(self): + b = self._number.to_bytes(6, byteorder="big") + return ':'.join(format(s, '02x') for s in b) + + def to_str_format(self): + b = self._number.to_bytes(6, byteorder="big") + return ''.join(format(s, '02x') for s in b) + + def create(self): + last_number = self.last_used_index() + + if last_number == int('0xffffffff', 16): + raise UncloudException("Exhausted all possible mac addresses - try to free some") + + next_number = last_number + 1 + self._number = self.prefix + next_number + + #next_number_string = "{:012x}".format(next_number) + #next_mac = self.int_to_mac(next_mac_number) + # db_entry = {} + # db_entry['vm_uuid'] = vmuuid + # db_entry['index'] = next_number + # db_entry['mac_address'] = next_mac + + # should be one transaction + # self.db.increment("last_used_index") + # self.db.set("used/{}".format(next_mac), + # db_entry, as_json=True) + + def __int__(self): + return self._number + + def __repr__(self): + return self.to_str_format() + + def __str__(self): + return self.to_colon_format() diff --git a/uncloud/hack/main.py b/uncloud/hack/main.py new file mode 100644 index 0000000..9607ec2 --- /dev/null +++ b/uncloud/hack/main.py @@ -0,0 +1,92 @@ +import argparse +import logging + +from uncloud.hack.vm import VM +from uncloud.hack.config import Config +from uncloud.hack.mac import MAC +from uncloud.hack.net import VXLANBridge, DNSRA + +from uncloud import UncloudException + +arg_parser = argparse.ArgumentParser('hack', add_help=False) + #description="Commands that are unfinished - use at own risk") +arg_parser.add_argument('--last-used-mac', action='store_true') +arg_parser.add_argument('--get-new-mac', action='store_true') + +arg_parser.add_argument('--init-network', help="Initialise networking", action='store_true') +arg_parser.add_argument('--create-vxlan', help="Initialise networking", action='store_true') +arg_parser.add_argument('--network', help="/64 IPv6 network") +arg_parser.add_argument('--vxlan-uplink-device', help="The VXLAN underlay device, i.e. eth0") +arg_parser.add_argument('--vni', help="VXLAN ID (decimal)", type=int) +arg_parser.add_argument('--run-dns-ra', action='store_true', + help="Provide router advertisements and DNS resolution via dnsmasq") +arg_parser.add_argument('--use-sudo', help="Use sudo for command requiring root!", action='store_true') + +arg_parser.add_argument('--create-vm', action='store_true') +arg_parser.add_argument('--destroy-vm', action='store_true') +arg_parser.add_argument('--get-vm-status', action='store_true') +arg_parser.add_argument('--get-vm-vnc', action='store_true') +arg_parser.add_argument('--list-vms', action='store_true') +arg_parser.add_argument('--memory', help="Size of memory (GB)", type=int) +arg_parser.add_argument('--cores', help="Amount of CPU cores", type=int) +arg_parser.add_argument('--image', help="Path (under hackprefix) to OS image") +arg_parser.add_argument('--uuid', help="VM UUID") + +arg_parser.add_argument('--no-db', help="Disable connection to etcd. For local testing only!", action='store_true') +arg_parser.add_argument('--hackprefix', help="hackprefix, if you need it you know it (it's where the iso is located and ifup/down.sh") + + +log = logging.getLogger(__name__) + + +def main(arguments): + config = Config(arguments) + + if arguments['create_vm']: + vm = VM(config) + vm.create() + + if arguments['destroy_vm']: + vm = VM(config) + vm.stop() + + if arguments['get_vm_status']: + vm = VM(config) + vm.status() + + if arguments['get_vm_vnc']: + vm = VM(config) + vm.vnc_addr() + + if arguments['list_vms']: + vm = VM(config) + vm.list() + + if arguments['last_used_mac']: + m = MAC(config) + print(m.last_used_mac()) + + if arguments['get_new_mac']: + print(MAC(config).get_next()) + + #if arguments['init_network']: + if arguments['create_vxlan']: + if not arguments['network'] or not arguments['vni'] or not arguments['vxlan_uplink_device']: + raise UncloudException("Initialising the network requires an IPv6 network and a VNI. You can use fd00::/64 and vni=1 for testing (non production!)") + vb = VXLANBridge(vni=arguments['vni'], + route=arguments['network'], + uplinkdev=arguments['vxlan_uplink_device'], + use_sudo=arguments['use_sudo']) + vb._setup_vxlan() + vb._setup_bridge() + vb._add_vxlan_to_bridge() + vb._route_network() + + if arguments['run_dns_ra']: + if not arguments['network'] or not arguments['vni']: + raise UncloudException("Providing DNS/RAs requires a /64 IPv6 network and a VNI. You can use fd00::/64 and vni=1 for testing (non production!)") + + dnsra = DNSRA(route=arguments['network'], + vni=arguments['vni'], + use_sudo=arguments['use_sudo']) + dnsra._setup_dnsmasq() diff --git a/uncloud/hack/net.py b/uncloud/hack/net.py new file mode 100644 index 0000000..f28ab7f --- /dev/null +++ b/uncloud/hack/net.py @@ -0,0 +1,116 @@ +import subprocess +import ipaddress +import logging + + +from uncloud import UncloudException + +log = logging.getLogger(__name__) + + +class VXLANBridge(object): + cmd_create_vxlan = "{sudo}ip -6 link add {vxlandev} type vxlan id {vni_dec} dstport 4789 group {multicast_address} dev {uplinkdev} ttl 5" + cmd_up_dev = "{sudo}ip link set {dev} up" + cmd_create_bridge="{sudo}ip link add {bridgedev} type bridge" + cmd_add_to_bridge="{sudo}ip link set {vxlandev} master {bridgedev} up" + cmd_add_addr="{sudo}ip addr add {ip} dev {bridgedev}" + cmd_add_route_dev="{sudo}ip route add {route} dev {bridgedev}" + + # VXLAN ids are at maximum 24 bit - use a /104 + multicast_network = ipaddress.IPv6Network("ff05::/104") + max_vni = (2**24)-1 + + def __init__(self, + vni, + uplinkdev, + route=None, + use_sudo=False): + self.config = {} + + if vni > self.max_vni: + raise UncloudException("VNI must be in the range of 0 .. {}".format(self.max_vni)) + + if use_sudo: + self.config['sudo'] = 'sudo ' + else: + self.config['sudo'] = '' + + self.config['vni_dec'] = vni + self.config['vni_hex'] = "{:x}".format(vni) + self.config['multicast_address'] = self.multicast_network[vni] + + self.config['route_network'] = ipaddress.IPv6Network(route) + self.config['route'] = route + + self.config['uplinkdev'] = uplinkdev + self.config['vxlandev'] = "vx{}".format(self.config['vni_hex']) + self.config['bridgedev'] = "br{}".format(self.config['vni_hex']) + + + def setup_networking(self): + pass + + def _setup_vxlan(self): + self._execute_cmd(self.cmd_create_vxlan) + self._execute_cmd(self.cmd_up_dev, dev=self.config['vxlandev']) + + def _setup_bridge(self): + self._execute_cmd(self.cmd_create_bridge) + self._execute_cmd(self.cmd_up_dev, dev=self.config['bridgedev']) + + def _route_network(self): + self._execute_cmd(self.cmd_add_route_dev) + + def _add_vxlan_to_bridge(self): + self._execute_cmd(self.cmd_add_to_bridge) + + def _execute_cmd(self, cmd_string, **kwargs): + cmd = cmd_string.format(**self.config, **kwargs) + log.info("Executing: {}".format(cmd)) + subprocess.run(cmd.split()) + +class ManagementBridge(VXLANBridge): + pass + + +class DNSRA(object): + # VXLAN ids are at maximum 24 bit + max_vni = (2**24)-1 + + + # Command to start dnsmasq + cmd_start_dnsmasq="{sudo}dnsmasq --interface={bridgedev} --bind-interfaces --dhcp-range={route},ra-only,infinite --enable-ra" + + def __init__(self, + vni, + route=None, + use_sudo=False): + self.config = {} + + if vni > self.max_vni: + raise UncloudException("VNI must be in the range of 0 .. {}".format(self.max_vni)) + + if use_sudo: + self.config['sudo'] = 'sudo ' + else: + self.config['sudo'] = '' + + #TODO: remove if not needed + #self.config['vni_dec'] = vni + self.config['vni_hex'] = "{:x}".format(vni) + + # dnsmasq only wants the network without the prefix, therefore, cut it off + self.config['route'] = ipaddress.IPv6Network(route).network_address + self.config['bridgedev'] = "br{}".format(self.config['vni_hex']) + + def _setup_dnsmasq(self): + self._execute_cmd(self.cmd_start_dnsmasq) + + def _execute_cmd(self, cmd_string, **kwargs): + cmd = cmd_string.format(**self.config, **kwargs) + log.info("Executing: {}".format(cmd)) + print("Executing: {}".format(cmd)) + subprocess.run(cmd.split()) + +class Firewall(object): + pass diff --git a/ucloud/hack/nftables.conf b/uncloud/hack/nftables.conf similarity index 100% rename from ucloud/hack/nftables.conf rename to uncloud/hack/nftables.conf diff --git a/ucloud/hack/rc-scripts/ucloud-api b/uncloud/hack/rc-scripts/ucloud-api similarity index 100% rename from ucloud/hack/rc-scripts/ucloud-api rename to uncloud/hack/rc-scripts/ucloud-api diff --git a/ucloud/hack/rc-scripts/ucloud-host b/uncloud/hack/rc-scripts/ucloud-host similarity index 100% rename from ucloud/hack/rc-scripts/ucloud-host rename to uncloud/hack/rc-scripts/ucloud-host diff --git a/ucloud/hack/rc-scripts/ucloud-metadata b/uncloud/hack/rc-scripts/ucloud-metadata similarity index 100% rename from ucloud/hack/rc-scripts/ucloud-metadata rename to uncloud/hack/rc-scripts/ucloud-metadata diff --git a/ucloud/hack/rc-scripts/ucloud-scheduler b/uncloud/hack/rc-scripts/ucloud-scheduler similarity index 100% rename from ucloud/hack/rc-scripts/ucloud-scheduler rename to uncloud/hack/rc-scripts/ucloud-scheduler diff --git a/uncloud/hack/uncloud-hack-init-host b/uncloud/hack/uncloud-hack-init-host new file mode 100644 index 0000000..787ff80 --- /dev/null +++ b/uncloud/hack/uncloud-hack-init-host @@ -0,0 +1,26 @@ +id=100 +rawdev=eth0 + +# create vxlan +ip -6 link add vxlan${id} type vxlan \ + id ${id} \ + dstport 4789 \ + group ff05::${id} \ + dev ${rawdev} \ + ttl 5 + +ip link set vxlan${id} up + +# create bridge +ip link set vxlan${id} up +ip link set br${id} up + +# Add vxlan into bridge +ip link set vxlan${id} master br${id} + + +# useradd -m uncloud +# [18:05] tablett.place10:~# id uncloud +# uid=1000(uncloud) gid=1000(uncloud) groups=1000(uncloud),34(kvm),36(qemu) +# apk add qemu-system-x86_64 +# also needs group netdev diff --git a/uncloud/hack/uncloud-run-vm b/uncloud/hack/uncloud-run-vm new file mode 100644 index 0000000..33e5860 --- /dev/null +++ b/uncloud/hack/uncloud-run-vm @@ -0,0 +1,25 @@ +#!/bin/sh + +if [ $# -ne 1 ]; then + echo $0 vmid + exit 1 +fi + +id=$1; shift + +memory=512 +macaddress=02:00:b9:cb:70:${id} +netname=net${id}-1 + +qemu-system-x86_64 \ + -name uncloud-${id} \ + -accel kvm \ + -m ${memory} \ + -smp 2,sockets=2,cores=1,threads=1 \ + -device virtio-net-pci,netdev=net0,mac=$macaddress \ + -netdev tap,id=net0,ifname=${netname},script=no,downscript=no \ + -vnc [::]:0 + +# To be changed: +# -vnc to unix path +# or -spice diff --git a/uncloud/hack/vm.py b/uncloud/hack/vm.py new file mode 100755 index 0000000..ac403d8 --- /dev/null +++ b/uncloud/hack/vm.py @@ -0,0 +1,136 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +# +# 2020 Nico Schottelius (nico.schottelius at ungleich.ch) +# +# This file is part of uncloud. +# +# uncloud is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# uncloud is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with uncloud. If not, see . + +# This module is directly called from the hack module, and can be used as follow: +# +# Create a new VM with default CPU/Memory. The path of the image file is relative to $hackprefix. +# `uncloud hack --hackprefix /tmp/hackcloud --create-vm --image mysuperimage.qcow2` +# +# List running VMs (returns a list of UUIDs). +# `uncloud hack --hackprefix /tmp/hackcloud --list-vms +# +# Get VM status: +# `uncloud hack --hackprefix /tmp/hackcloud --get-vm-status --uuid my-vm-uuid` +# +# Stop a VM: +# `uncloud hack --hackprefix /tmp/hackcloud --destroy-vm --uuid my-vm-uuid` +# `` + +import subprocess +import uuid +import os +import logging + +from uncloud.hack.db import DB +from uncloud.hack.mac import MAC +from uncloud.vmm import VMM + +log = logging.getLogger(__name__) +log.setLevel(logging.DEBUG) + +class VM(object): + def __init__(self, config): + self.config = config + + #TODO: Enable etcd lookup + self.no_db = self.config.arguments['no_db'] + if not self.no_db: + self.db = DB(self.config, prefix="/vm") + + # General CLI arguments. + self.hackprefix = self.config.arguments['hackprefix'] + self.uuid = self.config.arguments['uuid'] + self.memory = self.config.arguments['memory'] or '1024M' + self.cores = self.config.arguments['cores'] or 1 + if self.config.arguments['image']: + self.image = os.path.join(self.hackprefix, self.config.arguments['image']) + else: + self.image = None + + # External components. + self.vmm = VMM(vmm_backend=self.hackprefix) + self.mac = MAC(self.config) + + # Harcoded & generated values. + self.owner = 'uncoud' + self.image_format='qcow2' + self.accel = 'kvm' + self.threads = 1 + self.ifup = os.path.join(self.hackprefix, "ifup.sh") + self.ifdown = os.path.join(self.hackprefix, "ifdown.sh") + self.ifname = "uc{}".format(self.mac.to_str_format()) + + def get_qemu_args(self): + command = ( + "-name {owner}-{name}" + " -machine pc,accel={accel}" + " -drive file={image},format={image_format},if=virtio" + " -device virtio-rng-pci" + " -m {memory} -smp cores={cores},threads={threads}" + " -netdev tap,id=netmain,script={ifup},downscript={ifdown},ifname={ifname}" + " -device virtio-net-pci,netdev=netmain,id=net0,mac={mac}" + ).format( + owner=self.owner, name=self.uuid, + accel=self.accel, + image=self.image, image_format=self.image_format, + memory=self.memory, cores=self.cores, threads=self.threads, + ifup=self.ifup, ifdown=self.ifdown, ifname=self.ifname, + mac=self.mac + ) + + return command.split(" ") + + def create(self): + # New VM: new UUID, new MAC. + self.uuid = str(uuid.uuid4()) + self.mac.create() + + qemu_args = self.get_qemu_args() + log.debug("QEMU args passed to VMM: {}".format(qemu_args)) + self.vmm.start( + uuid=self.uuid, + migration=False, + *qemu_args + ) + + def stop(self): + if not self.uuid: + print("Please specific an UUID with the --uuid flag.") + exit(1) + + self.vmm.stop(self.uuid) + + def status(self): + if not self.uuid: + print("Please specific an UUID with the --uuid flag.") + exit(1) + + print(self.vmm.get_status(self.uuid)) + + def vnc_addr(self): + if not self.uuid: + print("Please specific an UUID with the --uuid flag.") + exit(1) + + print(self.vmm.get_vnc(self.uuid)) + + def list(self): + print(self.vmm.discover()) + diff --git a/ucloud/host/__init__.py b/uncloud/host/__init__.py similarity index 100% rename from ucloud/host/__init__.py rename to uncloud/host/__init__.py diff --git a/uncloud/host/main.py b/uncloud/host/main.py new file mode 100755 index 0000000..f680991 --- /dev/null +++ b/uncloud/host/main.py @@ -0,0 +1,123 @@ +import argparse +import multiprocessing as mp +import time + +from uuid import uuid4 + +from uncloud.common.request import RequestEntry, RequestType +from uncloud.common.shared import shared +from uncloud.common.vm import VMStatus +from uncloud.vmm import VMM +from os.path import join as join_path + +from . import virtualmachine, logger + +arg_parser = argparse.ArgumentParser('host', add_help=False) +arg_parser.add_argument('--hostname', required=True) + + +def update_heartbeat(hostname): + """Update Last HeartBeat Time for :param hostname: in etcd""" + host_pool = shared.host_pool + this_host = next( + filter(lambda h: h.hostname == hostname, host_pool.hosts), None + ) + while True: + this_host.update_heartbeat() + host_pool.put(this_host) + time.sleep(10) + + +def maintenance(host): + vmm = VMM() + running_vms = vmm.discover() + for vm_uuid in running_vms: + if vmm.is_running(vm_uuid) and vmm.get_status(vm_uuid) == 'running': + logger.debug('VM {} is running on {}'.format(vm_uuid, host)) + vm = shared.vm_pool.get( + join_path(shared.settings['etcd']['vm_prefix'], vm_uuid) + ) + vm.status = VMStatus.running + vm.vnc_socket = vmm.get_vnc(vm_uuid) + vm.hostname = host + shared.vm_pool.put(vm) + + +def main(arguments): + hostname = arguments['hostname'] + host_pool = shared.host_pool + host = next(filter(lambda h: h.hostname == hostname, host_pool.hosts), None) + + # Does not yet exist, create it + if not host: + host_key = join_path( + shared.settings['etcd']['host_prefix'], uuid4().hex + ) + host_entry = { + 'specs': '', + 'hostname': hostname, + 'status': 'DEAD', + 'last_heartbeat': '', + } + shared.etcd_client.put( + host_key, host_entry, value_in_json=True + ) + + # update, get ourselves now for sure + host = next(filter(lambda h: h.hostname == hostname, host_pool.hosts), None) + + try: + heartbeat_updating_process = mp.Process(target=update_heartbeat, args=(hostname,)) + heartbeat_updating_process.start() + except Exception as e: + raise Exception('uncloud-host heartbeat updating mechanism is not working') from e + + # The below while True is neccessary for gracefully handling leadership transfer and temporary + # unavailability in etcd. Why does it work? It works because the get_prefix,watch_prefix return + # iter([]) that is iterator of empty list on exception (that occur due to above mentioned reasons) + # which ends the loop immediately. So, having it inside infinite loop we try again and again to + # get prefix until either success or deamon death comes. + while True: + for events_iterator in [ + shared.etcd_client.get_prefix(shared.settings['etcd']['request_prefix'], value_in_json=True, + raise_exception=False), + shared.etcd_client.watch_prefix(shared.settings['etcd']['request_prefix'], value_in_json=True, + raise_exception=False) + ]: + for request_event in events_iterator: + request_event = RequestEntry(request_event) + + maintenance(host.key) + + if request_event.hostname == host.key: + logger.debug('VM Request: %s on Host %s', request_event, host.hostname) + + shared.request_pool.client.client.delete(request_event.key) + vm_entry = shared.etcd_client.get( + join_path(shared.settings['etcd']['vm_prefix'], request_event.uuid) + ) + + logger.debug('VM hostname: {}'.format(vm_entry.value)) + + vm = virtualmachine.VM(vm_entry) + if request_event.type == RequestType.StartVM: + vm.start() + + elif request_event.type == RequestType.StopVM: + vm.stop() + + elif request_event.type == RequestType.DeleteVM: + vm.delete() + + elif request_event.type == RequestType.InitVMMigration: + vm.start(destination_host_key=host.key) + + elif request_event.type == RequestType.TransferVM: + destination_host = host_pool.get(request_event.destination_host_key) + if destination_host: + vm.migrate( + destination_host=destination_host.hostname, + destination_sock_path=request_event.destination_sock_path, + ) + else: + logger.error('Host %s not found!', request_event.destination_host_key) diff --git a/uncloud/host/virtualmachine.py b/uncloud/host/virtualmachine.py new file mode 100755 index 0000000..a592efc --- /dev/null +++ b/uncloud/host/virtualmachine.py @@ -0,0 +1,303 @@ +# QEMU Manual +# https://qemu.weilnetz.de/doc/qemu-doc.html + +# For QEMU Monitor Protocol Commands Information, See +# https://qemu.weilnetz.de/doc/qemu-doc.html#pcsys_005fmonitor + +import os +import subprocess as sp +import ipaddress + +from string import Template +from os.path import join as join_path + +from uncloud.common.request import RequestEntry, RequestType +from uncloud.common.vm import VMStatus, declare_stopped +from uncloud.common.network import create_dev, delete_network_interface +from uncloud.common.schemas import VMSchema, NetworkSchema +from uncloud.host import logger +from uncloud.common.shared import shared +from uncloud.vmm import VMM + +from marshmallow import ValidationError + + +class VM: + def __init__(self, vm_entry): + self.schema = VMSchema() + self.vmm = VMM() + self.key = vm_entry.key + try: + self.vm = self.schema.loads(vm_entry.value) + except ValidationError: + logger.exception( + "Couldn't validate VM Entry", vm_entry.value + ) + self.vm = None + else: + self.uuid = vm_entry.key.split("/")[-1] + self.host_key = self.vm["hostname"] + logger.debug('VM Hostname {}'.format(self.host_key)) + + def get_qemu_args(self): + command = ( + "-drive file={file},format=raw,if=virtio" + " -device virtio-rng-pci" + " -m {memory} -smp cores={cores},threads={threads}" + " -name {owner}_{name}" + ).format( + owner=self.vm["owner"], + name=self.vm["name"], + memory=int(self.vm["specs"]["ram"].to_MB()), + cores=self.vm["specs"]["cpu"], + threads=1, + file=shared.storage_handler.qemu_path_string(self.uuid), + ) + + return command.split(" ") + + def start(self, destination_host_key=None): + migration = False + if destination_host_key: + migration = True + + self.create() + try: + network_args = self.create_network_dev() + except Exception as err: + declare_stopped(self.vm) + self.vm["log"].append("Cannot Setup Network Properly") + logger.error("Cannot Setup Network Properly for vm %s", self.uuid, exc_info=err) + else: + self.vmm.start( + uuid=self.uuid, + migration=migration, + *self.get_qemu_args(), + *network_args + ) + + status = self.vmm.get_status(self.uuid) + logger.debug('VM {} status is {}'.format(self.uuid, status)) + if status == "running": + self.vm["status"] = VMStatus.running + self.vm["vnc_socket"] = self.vmm.get_vnc(self.uuid) + elif status == "inmigrate": + r = RequestEntry.from_scratch( + type=RequestType.TransferVM, # Transfer VM + hostname=self.host_key, # Which VM should get this request. It is source host + uuid=self.uuid, # uuid of VM + destination_sock_path=join_path( + self.vmm.socket_dir, self.uuid + ), + destination_host_key=destination_host_key, # Where source host transfer VM + request_prefix=shared.settings["etcd"]["request_prefix"], + ) + shared.request_pool.put(r) + else: + self.stop() + declare_stopped(self.vm) + logger.debug('VM {} has hostname {}'.format(self.uuid, self.vm['hostname'])) + self.sync() + + def stop(self): + self.vmm.stop(self.uuid) + self.delete_network_dev() + declare_stopped(self.vm) + self.sync() + + def migrate(self, destination_host, destination_sock_path): + self.vmm.transfer( + src_uuid=self.uuid, + destination_sock_path=destination_sock_path, + host=destination_host, + ) + + def create_network_dev(self): + command = "" + for network_mac_and_tap in self.vm["network"]: + network_name, mac, tap = network_mac_and_tap + + _key = os.path.join( + shared.settings["etcd"]["network_prefix"], + self.vm["owner"], + network_name, + ) + network = shared.etcd_client.get(_key, value_in_json=True) + network_schema = NetworkSchema() + try: + network = network_schema.load(network.value) + except ValidationError: + continue + + if network["type"] == "vxlan": + tap = create_vxlan_br_tap( + _id=network["id"], + _dev=shared.settings["network"]["vxlan_phy_dev"], + tap_id=tap, + ip=network["ipv6"], + ) + + all_networks = shared.etcd_client.get_prefix( + shared.settings["etcd"]["network_prefix"], + value_in_json=True, + ) + + if ipaddress.ip_network(network["ipv6"]).is_global: + update_radvd_conf(all_networks) + + command += ( + "-netdev tap,id=vmnet{net_id},ifname={tap},script=no,downscript=no" + " -device virtio-net-pci,netdev=vmnet{net_id},mac={mac}".format( + tap=tap, net_id=network["id"], mac=mac + ) + ) + + if command: + command = command.split(' ') + + return command + + def delete_network_dev(self): + try: + for network in self.vm["network"]: + network_name = network[0] + _ = network[1] # tap_mac + tap_id = network[2] + + delete_network_interface("tap{}".format(tap_id)) + + owners_vms = shared.vm_pool.by_owner(self.vm["owner"]) + owners_running_vms = shared.vm_pool.by_status( + VMStatus.running, _vms=owners_vms + ) + + networks = map( + lambda n: n[0], + map(lambda vm: vm.network, owners_running_vms), + ) + networks_in_use_by_user_vms = [vm[0] for vm in networks] + if network_name not in networks_in_use_by_user_vms: + network_entry = resolve_network( + network[0], self.vm["owner"] + ) + if network_entry: + network_type = network_entry.value["type"] + network_id = network_entry.value["id"] + if network_type == "vxlan": + delete_network_interface( + "br{}".format(network_id) + ) + delete_network_interface( + "vxlan{}".format(network_id) + ) + except Exception: + logger.exception("Exception in network interface deletion") + + def create(self): + if shared.storage_handler.is_vm_image_exists(self.uuid): + # File Already exists. No Problem Continue + logger.debug("Image for vm %s exists", self.uuid) + else: + if shared.storage_handler.make_vm_image( + src=self.vm["image_uuid"], dest=self.uuid + ): + if not shared.storage_handler.resize_vm_image( + path=self.uuid, + size=int(self.vm["specs"]["os-ssd"].to_MB()), + ): + self.vm["status"] = VMStatus.error + else: + logger.info("New VM Created") + + def sync(self): + shared.etcd_client.put( + self.key, self.schema.dump(self.vm), value_in_json=True + ) + + def delete(self): + self.stop() + + if shared.storage_handler.is_vm_image_exists(self.uuid): + r_status = shared.storage_handler.delete_vm_image(self.uuid) + if r_status: + shared.etcd_client.client.delete(self.key) + else: + shared.etcd_client.client.delete(self.key) + + +def resolve_network(network_name, network_owner): + network = shared.etcd_client.get( + join_path( + shared.settings["etcd"]["network_prefix"], + network_owner, + network_name, + ), + value_in_json=True, + ) + return network + + +def create_vxlan_br_tap(_id, _dev, tap_id, ip=None): + network_script_base = os.path.join( + os.path.dirname(os.path.dirname(__file__)), "network" + ) + vxlan = create_dev( + script=os.path.join(network_script_base, "create-vxlan.sh"), + _id=_id, + dev=_dev, + ) + if vxlan: + bridge = create_dev( + script=os.path.join( + network_script_base, "create-bridge.sh" + ), + _id=_id, + dev=vxlan, + ip=ip, + ) + if bridge: + tap = create_dev( + script=os.path.join( + network_script_base, "create-tap.sh" + ), + _id=str(tap_id), + dev=bridge, + ) + if tap: + return tap + + +def update_radvd_conf(all_networks): + network_script_base = os.path.join( + os.path.dirname(os.path.dirname(__file__)), "network" + ) + + networks = { + net.value["ipv6"]: net.value["id"] + for net in all_networks + if net.value.get("ipv6") + and ipaddress.ip_network(net.value.get("ipv6")).is_global + } + radvd_template = open( + os.path.join(network_script_base, "radvd-template.conf"), "r" + ).read() + radvd_template = Template(radvd_template) + + content = [ + radvd_template.safe_substitute( + bridge="br{}".format(networks[net]), prefix=net + ) + for net in networks + if networks.get(net) + ] + with open("/etc/radvd.conf", "w") as radvd_conf: + radvd_conf.writelines(content) + try: + sp.check_output(["systemctl", "restart", "radvd"]) + except sp.CalledProcessError: + try: + sp.check_output(["service", "radvd", "restart"]) + except sp.CalledProcessError as err: + raise err.__class__( + "Cannot start/restart radvd service", err.cmd + ) from err diff --git a/ucloud/imagescanner/__init__.py b/uncloud/imagescanner/__init__.py similarity index 100% rename from ucloud/imagescanner/__init__.py rename to uncloud/imagescanner/__init__.py diff --git a/uncloud/imagescanner/main.py b/uncloud/imagescanner/main.py new file mode 100755 index 0000000..ee9da2e --- /dev/null +++ b/uncloud/imagescanner/main.py @@ -0,0 +1,121 @@ +import json +import os +import argparse +import subprocess as sp + +from os.path import join as join_path +from uncloud.common.shared import shared +from uncloud.imagescanner import logger + + +arg_parser = argparse.ArgumentParser('imagescanner', add_help=False) + + +def qemu_img_type(path): + qemu_img_info_command = [ + "qemu-img", + "info", + "--output", + "json", + path, + ] + try: + qemu_img_info = sp.check_output(qemu_img_info_command) + except Exception as e: + logger.exception(e) + return None + else: + qemu_img_info = json.loads(qemu_img_info.decode("utf-8")) + return qemu_img_info["format"] + + +def main(arguments): + # We want to get images entries that requests images to be created + images = shared.etcd_client.get_prefix( + shared.settings["etcd"]["image_prefix"], value_in_json=True + ) + images_to_be_created = list( + filter(lambda im: im.value["status"] == "TO_BE_CREATED", images) + ) + + for image in images_to_be_created: + try: + image_uuid = image.key.split("/")[-1] + image_owner = image.value["owner"] + image_filename = image.value["filename"] + image_store_name = image.value["store_name"] + image_full_path = join_path( + shared.settings["storage"]["file_dir"], + image_owner, + image_filename, + ) + + image_stores = shared.etcd_client.get_prefix( + shared.settings["etcd"]["image_store_prefix"], + value_in_json=True, + ) + user_image_store = next( + filter( + lambda s, store_name=image_store_name: s.value[ + "name" + ] + == store_name, + image_stores, + ) + ) + + image_store_pool = user_image_store.value["attributes"][ + "pool" + ] + + except Exception as e: + logger.exception(e) + else: + # At least our basic data is available + qemu_img_convert_command = [ + "qemu-img", + "convert", + "-f", + "qcow2", + "-O", + "raw", + image_full_path, + "image.raw", + ] + + if qemu_img_type(image_full_path) == "qcow2": + try: + # Convert .qcow2 to .raw + sp.check_output(qemu_img_convert_command,) + + except sp.CalledProcessError: + logger.exception( + "Image convertion from .qcow2 to .raw failed." + ) + else: + # Import and Protect + r_status = shared.storage_handler.import_image( + src="image.raw", dest=image_uuid, protect=True + ) + if r_status: + # Everything is successfully done + image.value["status"] = "CREATED" + shared.etcd_client.put( + image.key, json.dumps(image.value) + ) + finally: + try: + os.remove("image.raw") + except Exception: + pass + + else: + # The user provided image is either not found or of invalid format + image.value["status"] = "INVALID_IMAGE" + shared.etcd_client.put( + image.key, json.dumps(image.value) + ) + + +if __name__ == "__main__": + main() diff --git a/uncloud/metadata/__init__.py b/uncloud/metadata/__init__.py new file mode 100644 index 0000000..eea436a --- /dev/null +++ b/uncloud/metadata/__init__.py @@ -0,0 +1,3 @@ +import logging + +logger = logging.getLogger(__name__) diff --git a/uncloud/metadata/main.py b/uncloud/metadata/main.py new file mode 100644 index 0000000..374260e --- /dev/null +++ b/uncloud/metadata/main.py @@ -0,0 +1,95 @@ +import os +import argparse + +from flask import Flask, request +from flask_restful import Resource, Api +from werkzeug.exceptions import HTTPException + +from uncloud.common.shared import shared + +app = Flask(__name__) +api = Api(app) + +app.logger.handlers.clear() + +DEFAULT_PORT=1234 + +arg_parser = argparse.ArgumentParser('metadata', add_help=False) +arg_parser.add_argument('--port', '-p', default=DEFAULT_PORT, help='By default bind to port {}'.format(DEFAULT_PORT)) + + +@app.errorhandler(Exception) +def handle_exception(e): + app.logger.error(e) + # pass through HTTP errors + if isinstance(e, HTTPException): + return e + + # now you're handling non-HTTP exceptions only + return {"message": "Server Error"}, 500 + + +def get_vm_entry(mac_addr): + return next( + filter( + lambda vm: mac_addr in list(zip(*vm.network))[1], + shared.vm_pool.vms, + ), + None, + ) + + +# https://stackoverflow.com/questions/37140846/how-to-convert-ipv6-link-local-address-to-mac-address-in-python +def ipv62mac(ipv6): + # remove subnet info if given + subnet_index = ipv6.find("/") + if subnet_index != -1: + ipv6 = ipv6[:subnet_index] + + ipv6_parts = ipv6.split(":") + mac_parts = list() + for ipv6_part in ipv6_parts[-4:]: + while len(ipv6_part) < 4: + ipv6_part = "0" + ipv6_part + mac_parts.append(ipv6_part[:2]) + mac_parts.append(ipv6_part[-2:]) + + # modify parts to match MAC value + mac_parts[0] = "%02x" % (int(mac_parts[0], 16) ^ 2) + del mac_parts[4] + del mac_parts[3] + return ":".join(mac_parts) + + +class Root(Resource): + @staticmethod + def get(): + data = get_vm_entry(ipv62mac(request.remote_addr)) + + if not data: + return ( + {"message": "Metadata for such VM does not exists."}, + 404, + ) + else: + etcd_key = os.path.join( + shared.settings["etcd"]["user_prefix"], + data.value["owner_realm"], + data.value["owner"], + "key", + ) + etcd_entry = shared.etcd_client.get_prefix( + etcd_key, value_in_json=True + ) + user_personal_ssh_keys = [key.value for key in etcd_entry] + data.value["metadata"]["ssh-keys"] += user_personal_ssh_keys + return data.value["metadata"], 200 + + +api.add_resource(Root, "/") + + +def main(arguments): + port = arguments['port'] + debug = arguments['debug'] + app.run(debug=debug, host="::", port=port) diff --git a/ucloud/network/README b/uncloud/network/README similarity index 100% rename from ucloud/network/README rename to uncloud/network/README diff --git a/ucloud/scheduler/tests/__init__.py b/uncloud/network/__init__.py similarity index 100% rename from ucloud/scheduler/tests/__init__.py rename to uncloud/network/__init__.py diff --git a/ucloud/network/create-bridge.sh b/uncloud/network/create-bridge.sh similarity index 100% rename from ucloud/network/create-bridge.sh rename to uncloud/network/create-bridge.sh diff --git a/ucloud/network/create-tap.sh b/uncloud/network/create-tap.sh similarity index 100% rename from ucloud/network/create-tap.sh rename to uncloud/network/create-tap.sh diff --git a/ucloud/network/create-vxlan.sh b/uncloud/network/create-vxlan.sh similarity index 100% rename from ucloud/network/create-vxlan.sh rename to uncloud/network/create-vxlan.sh diff --git a/ucloud/network/radvd-template.conf b/uncloud/network/radvd-template.conf similarity index 100% rename from ucloud/network/radvd-template.conf rename to uncloud/network/radvd-template.conf diff --git a/uncloud/oneshot/__init__.py b/uncloud/oneshot/__init__.py new file mode 100644 index 0000000..eea436a --- /dev/null +++ b/uncloud/oneshot/__init__.py @@ -0,0 +1,3 @@ +import logging + +logger = logging.getLogger(__name__) diff --git a/uncloud/oneshot/main.py b/uncloud/oneshot/main.py new file mode 100644 index 0000000..5b9b61c --- /dev/null +++ b/uncloud/oneshot/main.py @@ -0,0 +1,123 @@ +import argparse +import os + + +from pathlib import Path +from uncloud.vmm import VMM +from uncloud.host.virtualmachine import update_radvd_conf, create_vxlan_br_tap + +from . import virtualmachine, logger + +### +# Argument parser loaded by scripts/uncloud. +arg_parser = argparse.ArgumentParser('oneshot', add_help=False) + +# Actions. +arg_parser.add_argument('--list', action='store_true', + help='list UUID and name of running VMs') +arg_parser.add_argument('--start', nargs=3, + metavar=('IMAGE', 'UPSTREAM_INTERFACE', 'NETWORK'), + help='start a VM using the OS IMAGE (full path), configuring networking on NETWORK IPv6 prefix') +arg_parser.add_argument('--stop', metavar='UUID', + help='stop a VM') +arg_parser.add_argument('--get-status', metavar='UUID', + help='return the status of the VM') +arg_parser.add_argument('--get-vnc', metavar='UUID', + help='return the path of the VNC socket of the VM') +arg_parser.add_argument('--reconfigure-radvd', metavar='NETWORK', + help='regenerate and reload RADVD configuration for NETWORK IPv6 prefix') + +# Arguments. +arg_parser.add_argument('--workdir', default=Path.home(), + help='Working directory, defaulting to $HOME') +arg_parser.add_argument('--mac', + help='MAC address of the VM to create (--start)') +arg_parser.add_argument('--memory', type=int, + help='Memory (MB) to allocate (--start)') +arg_parser.add_argument('--cores', type=int, + help='Number of cores to allocate (--start)') +arg_parser.add_argument('--threads', type=int, + help='Number of threads to allocate (--start)') +arg_parser.add_argument('--image-format', choices=['raw', 'qcow2'], + help='Format of OS image (--start)') +arg_parser.add_argument('--accel', choices=['kvm', 'tcg'], default='kvm', + help='QEMU acceleration to use (--start)') +arg_parser.add_argument('--upstream-interface', default='eth0', + help='Name of upstream interface (--start)') + +### +# Helpers. + +# XXX: check if it is possible to use the type returned by ETCD queries. +class UncloudEntryWrapper: + def __init__(self, value): + self.value = value + + def value(self): + return self.value + +def status_line(vm): + return "VM: {} {} {}".format(vm.get_uuid(), vm.get_name(), vm.get_status()) + +### +# Entrypoint. + +def main(arguments): + # Initialize VMM. + workdir = arguments['workdir'] + vmm = VMM(vmm_backend=workdir) + + # Harcoded debug values. + net_id = 0 + + # Build VM configuration. + vm_config = {} + vm_options = [ + 'mac', 'memory', 'cores', 'threads', 'image', 'image_format', + '--upstream_interface', 'upstream_interface', 'network', 'accel' + ] + for option in vm_options: + if arguments.get(option): + vm_config[option] = arguments[option] + + vm_config['net_id'] = net_id + + # Execute requested VM action. + if arguments['reconfigure_radvd']: + # TODO: check that RADVD is available. + prefix = arguments['reconfigure_radvd'] + network = UncloudEntryWrapper({ + 'id': net_id, + 'ipv6': prefix + }) + + # Make use of uncloud.host.virtualmachine for network configuration. + update_radvd_conf([network]) + elif arguments['start']: + # Extract from --start positional arguments. Quite fragile. + vm_config['image'] = arguments['start'][0] + vm_config['network'] = arguments['start'][1] + vm_config['upstream_interface'] = arguments['start'][2] + + vm_config['tap_interface'] = "uc{}".format(len(vmm.discover())) + vm = virtualmachine.VM(vmm, vm_config) + vm.start() + elif arguments['stop']: + vm = virtualmachine.VM(vmm, {'uuid': arguments['stop']}) + vm = virtualmachine.VM(vmm, vm_config) + vm.stop() + elif arguments['get_status']: + vm = virtualmachine.VM(vmm, {'uuid': arguments['get_status']}) + print(status_line(vm)) + elif arguments['get_vnc']: + vm = virtualmachine.VM(vmm, {'uuid': arguments['get_vnc']}) + print(vm.get_vnc_addr()) + elif arguments['list']: + vms = vmm.discover() + print("Found {} VMs.".format(len(vms))) + for uuid in vms: + vm = virtualmachine.VM(vmm, {'uuid': uuid}) + print(status_line(vm)) + else: + print('Please specify an action: --start, --stop, --list,\ +--get-status, --get-vnc, --reconfigure-radvd') diff --git a/uncloud/oneshot/virtualmachine.py b/uncloud/oneshot/virtualmachine.py new file mode 100644 index 0000000..c8c2909 --- /dev/null +++ b/uncloud/oneshot/virtualmachine.py @@ -0,0 +1,81 @@ +import uuid +import os + +from uncloud.host.virtualmachine import create_vxlan_br_tap +from uncloud.oneshot import logger + +class VM(object): + def __init__(self, vmm, config): + self.config = config + self.vmm = vmm + + # Extract VM specs/metadata from configuration. + self.name = config.get('name', 'no-name') + self.memory = config.get('memory', 1024) + self.cores = config.get('cores', 1) + self.threads = config.get('threads', 1) + self.image_format = config.get('image_format', 'qcow2') + self.image = config.get('image') + self.uuid = config.get('uuid', str(uuid.uuid4())) + self.mac = config.get('mac') + self.accel = config.get('accel', 'kvm') + + self.net_id = config.get('net_id', 0) + self.upstream_interface = config.get('upstream_interface', 'eth0') + self.tap_interface = config.get('tap_interface', 'uc0') + self.network = config.get('network') + + def get_qemu_args(self): + command = ( + "-uuid {uuid} -name {name} -machine pc,accel={accel}" + " -drive file={image},format={image_format},if=virtio" + " -device virtio-rng-pci" + " -m {memory} -smp cores={cores},threads={threads}" + " -netdev tap,id=vmnet{net_id},ifname={tap},script=no,downscript=no" + " -device virtio-net-pci,netdev=vmnet{net_id},mac={mac}" + ).format( + uuid=self.uuid, name=self.name, accel=self.accel, + image=self.image, image_format=self.image_format, + memory=self.memory, cores=self.cores, threads=self.threads, + net_id=self.net_id, tap=self.tap_interface, mac=self.mac + ) + + return command.split(" ") + + def start(self): + # Check that VM image is available. + if not os.path.isfile(self.image): + logger.error("Image {} does not exist. Aborting.".format(self.image)) + + # Create Bridge, VXLAN and tap interface for VM. + create_vxlan_br_tap( + self.net_id, self.upstream_interface, self.tap_interface, self.network + ) + + # Generate config for and run QEMU. + qemu_args = self.get_qemu_args() + logger.debug("QEMU args for VM {}: {}".format(self.uuid, qemu_args)) + self.vmm.start( + uuid=self.uuid, + migration=False, + *qemu_args + ) + + def stop(self): + self.vmm.stop(self.uuid) + + def get_status(self): + return self.vmm.get_status(self.uuid) + + def get_vnc_addr(self): + return self.vmm.get_vnc(self.uuid) + + def get_uuid(self): + return self.uuid + + def get_name(self): + success, json = self.vmm.execute_command(uuid, 'query-name') + if success: + return json['return']['name'] + + return None diff --git a/uncloud/scheduler/__init__.py b/uncloud/scheduler/__init__.py new file mode 100644 index 0000000..eea436a --- /dev/null +++ b/uncloud/scheduler/__init__.py @@ -0,0 +1,3 @@ +import logging + +logger = logging.getLogger(__name__) diff --git a/ucloud/scheduler/helper.py b/uncloud/scheduler/helper.py similarity index 50% rename from ucloud/scheduler/helper.py rename to uncloud/scheduler/helper.py index ba577d6..79db322 100755 --- a/ucloud/scheduler/helper.py +++ b/uncloud/scheduler/helper.py @@ -3,10 +3,10 @@ from functools import reduce import bitmath -from ucloud.common.host import HostStatus -from ucloud.common.request import RequestEntry, RequestType -from ucloud.common.vm import VMStatus -from ucloud.config import vm_pool, host_pool, request_pool, env_vars +from uncloud.common.host import HostStatus +from uncloud.common.request import RequestEntry, RequestType +from uncloud.common.vm import VMStatus +from uncloud.common.shared import shared def accumulated_specs(vms_specs): @@ -23,17 +23,35 @@ def remaining_resources(host_specs, vms_specs): for component in _vms_specs: if isinstance(_vms_specs[component], str): - _vms_specs[component] = int(bitmath.parse_string_unsafe(_vms_specs[component]).to_MB()) + _vms_specs[component] = int( + bitmath.parse_string_unsafe( + _vms_specs[component] + ).to_MB() + ) elif isinstance(_vms_specs[component], list): - _vms_specs[component] = map(lambda x: int(bitmath.parse_string_unsafe(x).to_MB()), _vms_specs[component]) - _vms_specs[component] = reduce(lambda x, y: x + y, _vms_specs[component], 0) + _vms_specs[component] = map( + lambda x: int(bitmath.parse_string_unsafe(x).to_MB()), + _vms_specs[component], + ) + _vms_specs[component] = reduce( + lambda x, y: x + y, _vms_specs[component], 0 + ) for component in _remaining: if isinstance(_remaining[component], str): - _remaining[component] = int(bitmath.parse_string_unsafe(_remaining[component]).to_MB()) + _remaining[component] = int( + bitmath.parse_string_unsafe( + _remaining[component] + ).to_MB() + ) elif isinstance(_remaining[component], list): - _remaining[component] = map(lambda x: int(bitmath.parse_string_unsafe(x).to_MB()), _remaining[component]) - _remaining[component] = reduce(lambda x, y: x + y, _remaining[component], 0) + _remaining[component] = map( + lambda x: int(bitmath.parse_string_unsafe(x).to_MB()), + _remaining[component], + ) + _remaining[component] = reduce( + lambda x, y: x + y, _remaining[component], 0 + ) _remaining.subtract(_vms_specs) @@ -46,23 +64,27 @@ class NoSuitableHostFound(Exception): def get_suitable_host(vm_specs, hosts=None): if hosts is None: - hosts = host_pool.by_status(HostStatus.alive) + hosts = shared.host_pool.by_status(HostStatus.alive) for host in hosts: # Filter them by host_name - vms = vm_pool.by_host(host.key) + vms = shared.vm_pool.by_host(host.key) # Filter them by status - vms = vm_pool.by_status(VMStatus.running, vms) + vms = shared.vm_pool.by_status(VMStatus.running, vms) running_vms_specs = [vm.specs for vm in vms] # Accumulate all of their combined specs - running_vms_accumulated_specs = accumulated_specs(running_vms_specs) + running_vms_accumulated_specs = accumulated_specs( + running_vms_specs + ) # Find out remaining resources after # host_specs - already running vm_specs - remaining = remaining_resources(host.specs, running_vms_accumulated_specs) + remaining = remaining_resources( + host.specs, running_vms_accumulated_specs + ) # Find out remaining - new_vm_specs remaining = remaining_resources(remaining, vm_specs) @@ -75,7 +97,7 @@ def get_suitable_host(vm_specs, hosts=None): def dead_host_detection(): # Bring out your dead! - Monty Python and the Holy Grail - hosts = host_pool.by_status(HostStatus.alive) + hosts = shared.host_pool.by_status(HostStatus.alive) dead_hosts_keys = [] for host in hosts: @@ -89,25 +111,27 @@ def dead_host_detection(): def dead_host_mitigation(dead_hosts_keys): for host_key in dead_hosts_keys: - host = host_pool.get(host_key) + host = shared.host_pool.get(host_key) host.declare_dead() - vms_hosted_on_dead_host = vm_pool.by_host(host_key) + vms_hosted_on_dead_host = shared.vm_pool.by_host(host_key) for vm in vms_hosted_on_dead_host: - vm.declare_killed() - vm_pool.put(vm) - host_pool.put(host) + vm.status = "UNKNOWN" + shared.vm_pool.put(vm) + shared.host_pool.put(host) def assign_host(vm): vm.hostname = get_suitable_host(vm.specs) - vm_pool.put(vm) + shared.vm_pool.put(vm) - r = RequestEntry.from_scratch(type=RequestType.StartVM, - uuid=vm.uuid, - hostname=vm.hostname, - request_prefix=env_vars.get("REQUEST_PREFIX")) - request_pool.put(r) + r = RequestEntry.from_scratch( + type=RequestType.StartVM, + uuid=vm.uuid, + hostname=vm.hostname, + request_prefix=shared.settings["etcd"]["request_prefix"], + ) + shared.request_pool.put(r) vm.log.append("VM scheduled for starting") return vm.hostname diff --git a/uncloud/scheduler/main.py b/uncloud/scheduler/main.py new file mode 100755 index 0000000..38c07bf --- /dev/null +++ b/uncloud/scheduler/main.py @@ -0,0 +1,51 @@ +# TODO +# 1. send an email to an email address defined by env['admin-email'] +# if resources are finished +# 2. Introduce a status endpoint of the scheduler - +# maybe expose a prometheus compatible output + +import argparse + +from uncloud.common.request import RequestEntry, RequestType +from uncloud.common.shared import shared +from uncloud.scheduler import logger +from uncloud.scheduler.helper import (dead_host_mitigation, dead_host_detection, + assign_host, NoSuitableHostFound) + +arg_parser = argparse.ArgumentParser('scheduler', add_help=False) + + +def main(arguments): + # The below while True is neccessary for gracefully handling leadership transfer and temporary + # unavailability in etcd. Why does it work? It works because the get_prefix,watch_prefix return + # iter([]) that is iterator of empty list on exception (that occur due to above mentioned reasons) + # which ends the loop immediately. So, having it inside infinite loop we try again and again to + # get prefix until either success or deamon death comes. + while True: + for request_iterator in [ + shared.etcd_client.get_prefix(shared.settings['etcd']['request_prefix'], value_in_json=True, + raise_exception=False), + shared.etcd_client.watch_prefix(shared.settings['etcd']['request_prefix'], value_in_json=True, + raise_exception=False), + ]: + for request_event in request_iterator: + dead_host_mitigation(dead_host_detection()) + request_entry = RequestEntry(request_event) + + if request_entry.type == RequestType.ScheduleVM: + logger.debug('%s, %s', request_entry.key, request_entry.value) + + vm_entry = shared.vm_pool.get(request_entry.uuid) + if vm_entry is None: + logger.info('Trying to act on {} but it is deleted'.format(request_entry.uuid)) + continue + + shared.etcd_client.client.delete(request_entry.key) # consume Request + + try: + assign_host(vm_entry) + except NoSuitableHostFound: + vm_entry.add_log('Can\'t schedule VM. No Resource Left.') + shared.vm_pool.put(vm_entry) + + logger.info('No Resource Left. Emailing admin....') diff --git a/uncloud/scheduler/tests/__init__.py b/uncloud/scheduler/tests/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/ucloud/scheduler/tests/test_basics.py b/uncloud/scheduler/tests/test_basics.py similarity index 83% rename from ucloud/scheduler/tests/test_basics.py rename to uncloud/scheduler/tests/test_basics.py index 92b3a83..defeb23 100755 --- a/ucloud/scheduler/tests/test_basics.py +++ b/uncloud/scheduler/tests/test_basics.py @@ -15,7 +15,7 @@ from main import ( main, ) -from ucloud.config import etcd_client +from uncloud.config import etcd_client class TestFunctions(unittest.TestCase): @@ -70,9 +70,15 @@ class TestFunctions(unittest.TestCase): "last_heartbeat": datetime.utcnow().isoformat(), } with self.client.client.lock("lock"): - self.client.put(f"{self.host_prefix}/1", host1, value_in_json=True) - self.client.put(f"{self.host_prefix}/2", host2, value_in_json=True) - self.client.put(f"{self.host_prefix}/3", host3, value_in_json=True) + self.client.put( + f"{self.host_prefix}/1", host1, value_in_json=True + ) + self.client.put( + f"{self.host_prefix}/2", host2, value_in_json=True + ) + self.client.put( + f"{self.host_prefix}/3", host3, value_in_json=True + ) def create_vms(self): vm1 = json.dumps( @@ -146,15 +152,17 @@ class TestFunctions(unittest.TestCase): {"cpu": 8, "ram": 32}, ] self.assertEqual( - accumulated_specs(vms), {"ssd": 10, "cpu": 16, "ram": 48, "hdd": 10} + accumulated_specs(vms), + {"ssd": 10, "cpu": 16, "ram": 48, "hdd": 10}, ) def test_remaining_resources(self): host_specs = {"ssd": 10, "cpu": 16, "ram": 48, "hdd": 10} vms_specs = {"ssd": 10, "cpu": 32, "ram": 12, "hdd": 0} resultant_specs = {"ssd": 0, "cpu": -16, "ram": 36, "hdd": 10} - self.assertEqual(remaining_resources(host_specs, vms_specs), - resultant_specs) + self.assertEqual( + remaining_resources(host_specs, vms_specs), resultant_specs + ) def test_vmpool(self): self.p.join(1) @@ -167,7 +175,12 @@ class TestFunctions(unittest.TestCase): f"{self.vm_prefix}/1", { "owner": "meow", - "specs": {"cpu": 4, "ram": 8, "hdd": 100, "sdd": 256}, + "specs": { + "cpu": 4, + "ram": 8, + "hdd": 100, + "sdd": 256, + }, "hostname": f"{self.host_prefix}/3", "status": "SCHEDULED_DEPLOY", }, @@ -182,7 +195,12 @@ class TestFunctions(unittest.TestCase): f"{self.vm_prefix}/7", { "owner": "meow", - "specs": {"cpu": 10, "ram": 22, "hdd": 146, "sdd": 0}, + "specs": { + "cpu": 10, + "ram": 22, + "hdd": 146, + "sdd": 0, + }, "hostname": "", "status": "REQUESTED_NEW", }, @@ -197,7 +215,12 @@ class TestFunctions(unittest.TestCase): f"{self.vm_prefix}/7", { "owner": "meow", - "specs": {"cpu": 10, "ram": 22, "hdd": 146, "sdd": 0}, + "specs": { + "cpu": 10, + "ram": 22, + "hdd": 146, + "sdd": 0, + }, "hostname": "", "status": "REQUESTED_NEW", }, diff --git a/ucloud/scheduler/tests/test_dead_host_mechanism.py b/uncloud/scheduler/tests/test_dead_host_mechanism.py similarity index 70% rename from ucloud/scheduler/tests/test_dead_host_mechanism.py rename to uncloud/scheduler/tests/test_dead_host_mechanism.py index 0b403ef..466b9ee 100755 --- a/ucloud/scheduler/tests/test_dead_host_mechanism.py +++ b/uncloud/scheduler/tests/test_dead_host_mechanism.py @@ -6,11 +6,7 @@ from os.path import dirname BASE_DIR = dirname(dirname(__file__)) sys.path.insert(0, BASE_DIR) -from main import ( - dead_host_detection, - dead_host_mitigation, - config -) +from main import dead_host_detection, dead_host_mitigation, config class TestDeadHostMechanism(unittest.TestCase): @@ -52,13 +48,23 @@ class TestDeadHostMechanism(unittest.TestCase): "last_heartbeat": datetime(2011, 1, 1).isoformat(), } with self.client.client.lock("lock"): - self.client.put(f"{self.host_prefix}/1", host1, value_in_json=True) - self.client.put(f"{self.host_prefix}/2", host2, value_in_json=True) - self.client.put(f"{self.host_prefix}/3", host3, value_in_json=True) - self.client.put(f"{self.host_prefix}/4", host4, value_in_json=True) + self.client.put( + f"{self.host_prefix}/1", host1, value_in_json=True + ) + self.client.put( + f"{self.host_prefix}/2", host2, value_in_json=True + ) + self.client.put( + f"{self.host_prefix}/3", host3, value_in_json=True + ) + self.client.put( + f"{self.host_prefix}/4", host4, value_in_json=True + ) def test_dead_host_detection(self): - hosts = self.client.get_prefix(self.host_prefix, value_in_json=True) + hosts = self.client.get_prefix( + self.host_prefix, value_in_json=True + ) deads = dead_host_detection(hosts) self.assertEqual(deads, ["/test/host/2", "/test/host/3"]) return deads @@ -66,7 +72,9 @@ class TestDeadHostMechanism(unittest.TestCase): def test_dead_host_mitigation(self): deads = self.test_dead_host_detection() dead_host_mitigation(self.client, deads) - hosts = self.client.get_prefix(self.host_prefix, value_in_json=True) + hosts = self.client.get_prefix( + self.host_prefix, value_in_json=True + ) deads = dead_host_detection(hosts) self.assertEqual(deads, []) diff --git a/uncloud/vmm/__init__.py b/uncloud/vmm/__init__.py new file mode 100644 index 0000000..6db61eb --- /dev/null +++ b/uncloud/vmm/__init__.py @@ -0,0 +1,284 @@ +import os +import subprocess as sp +import logging +import socket +import json +import tempfile +import time + +from contextlib import suppress +from multiprocessing import Process +from os.path import join as join_path +from os.path import isdir + +logger = logging.getLogger(__name__) + + +class VMQMPHandles: + def __init__(self, path): + self.path = path + self.sock = socket.socket(socket.AF_UNIX) + self.file = self.sock.makefile() + + def __enter__(self): + self.sock.connect(self.path) + + # eat qmp greetings + self.file.readline() + + # init qmp + self.sock.sendall(b'{ "execute": "qmp_capabilities" }') + self.file.readline() + + return self.sock, self.file + + def __exit__(self, exc_type, exc_val, exc_tb): + self.file.close() + self.sock.close() + + if exc_type: + logger.error( + "Couldn't get handle for VM.", exc_type, exc_val, exc_tb + ) + raise exc_type("Couldn't get handle for VM.") from exc_type + + +class TransferVM(Process): + def __init__(self, src_uuid, dest_sock_path, host, socket_dir): + self.src_uuid = src_uuid + self.host = host + self.src_sock_path = os.path.join(socket_dir, self.src_uuid) + self.dest_sock_path = dest_sock_path + + super().__init__() + + def run(self): + with suppress(FileNotFoundError): + os.remove(self.src_sock_path) + + command = [ + "ssh", + "-nNT", + "-L", + "{}:{}".format(self.src_sock_path, self.dest_sock_path), + "root@{}".format(self.host), + ] + + try: + p = sp.Popen(command) + except Exception as e: + logger.error( + "Couldn' forward unix socks over ssh.", exc_info=e + ) + else: + time.sleep(2) + vmm = VMM() + logger.debug("Executing: ssh forwarding command: %s", command) + vmm.execute_command( + self.src_uuid, + command="migrate", + arguments={"uri": "unix:{}".format(self.src_sock_path)}, + ) + + while p.poll() is None: + success, output = vmm.execute_command(self.src_uuid, command="query-migrate") + if success: + status = output["return"]["status"] + logger.info('Migration Status: {}'.format(status)) + if status == "completed": + vmm.stop(self.src_uuid) + return + elif status in ['failed', 'cancelled']: + return + else: + logger.error("Couldn't be able to query VM {} that was in migration".format(self.src_uuid)) + return + + time.sleep(2) + + +class VMM: + # Virtual Machine Manager + def __init__( + self, + qemu_path="/usr/bin/qemu-system-x86_64", + vmm_backend=os.path.expanduser("~/uncloud/vmm/"), + ): + self.qemu_path = qemu_path + self.vmm_backend = vmm_backend + self.socket_dir = os.path.join(self.vmm_backend, "sock") + + if not os.path.isdir(self.vmm_backend): + logger.info( + "{} does not exists. Creating it...".format( + self.vmm_backend + ) + ) + os.makedirs(self.vmm_backend, exist_ok=True) + + if not os.path.isdir(self.socket_dir): + logger.info( + "{} does not exists. Creating it...".format( + self.socket_dir + ) + ) + os.makedirs(self.socket_dir, exist_ok=True) + + def is_running(self, uuid): + sock_path = os.path.join(self.socket_dir, uuid) + try: + sock = socket.socket(socket.AF_UNIX) + sock.connect(sock_path) + recv = sock.recv(4096) + except Exception as err: + # unix sock doesn't exists or it is closed + logger.debug( + "VM {} sock either don' exists or it is closed. It mean VM is stopped.".format( + uuid + ), + exc_info=err, + ) + else: + # if we receive greetings from qmp it mean VM is running + if len(recv) > 0: + return True + + with suppress(FileNotFoundError): + os.remove(sock_path) + + return False + + def start(self, *args, uuid, migration=False): + # start --> sucess? + migration_args = () + if migration: + migration_args = ( + "-incoming", + "unix:{}".format(os.path.join(self.socket_dir, uuid)), + ) + + if self.is_running(uuid): + logger.warning("Cannot start VM. It is already running.") + else: + qmp_arg = ( + "-qmp", + "unix:{},server,nowait".format( + join_path(self.socket_dir, uuid) + ), + ) + vnc_arg = ( + "-vnc", + "unix:{}".format(tempfile.NamedTemporaryFile().name), + ) + + command = [ + "sudo", + "-p", + "Enter password to start VM {}: ".format(uuid), + self.qemu_path, + *args, + *qmp_arg, + *migration_args, + *vnc_arg, + "-daemonize", + ] + try: + sp.check_output(command, stderr=sp.PIPE) + except sp.CalledProcessError as err: + logger.exception( + "Error occurred while starting VM.\nDetail %s", + err.stderr.decode("utf-8"), + ) + else: + sp.check_output( + ["sudo", "-p", "Enter password to correct permission for uncloud-vmm's directory", + "chmod", "-R", "o=rwx,g=rwx", self.vmm_backend] + ) + + # TODO: Find some good way to check whether the virtual machine is up and + # running without relying on non-guarenteed ways. + for _ in range(10): + time.sleep(2) + status = self.get_status(uuid) + if status in ["running", "inmigrate"]: + return status + logger.warning( + "Timeout on VM's status. Shutting down VM %s", uuid + ) + self.stop(uuid) + # TODO: What should we do more. VM can still continue to run in background. + # If we have pid of vm we can kill it using OS. + + def execute_command(self, uuid, command, **kwargs): + # execute_command -> sucess?, output + try: + with VMQMPHandles(os.path.join(self.socket_dir, uuid)) as ( + sock_handle, + file_handle, + ): + command_to_execute = {"execute": command, **kwargs} + sock_handle.sendall( + json.dumps(command_to_execute).encode("utf-8") + ) + output = file_handle.readline() + except Exception: + logger.exception( + "Error occurred while executing command and getting valid output from qmp" + ) + else: + try: + output = json.loads(output) + except Exception: + logger.exception( + "QMP Output isn't valid JSON. %s", output + ) + else: + return "return" in output, output + return False, None + + def stop(self, uuid): + success, output = self.execute_command( + command="quit", uuid=uuid + ) + return success + + def get_status(self, uuid): + success, output = self.execute_command( + command="query-status", uuid=uuid + ) + if success: + return output["return"]["status"] + else: + # TODO: Think about this for a little more + return "STOPPED" + + def discover(self): + vms = [ + uuid + for uuid in os.listdir(self.socket_dir) + if not isdir(join_path(self.socket_dir, uuid)) + ] + return vms + + def get_vnc(self, uuid): + success, output = self.execute_command( + uuid, command="query-vnc" + ) + if success: + return output["return"]["service"] + return None + + def transfer(self, src_uuid, destination_sock_path, host): + p = TransferVM( + src_uuid, + destination_sock_path, + socket_dir=self.socket_dir, + host=host, + ) + p.start() + + # TODO: the following method should clean things that went wrong + # e.g If VM migration fails or didn't start for long time + # i.e 15 minutes we should stop the waiting VM. + def maintenace(self): + pass