uncloud-mravi/ucloud/host/virtualmachine.py
meow 972bb5a920 - Better error reporting.
- Flask now uses application's logger instead of its own.
- ucloud file scanner refactored.
2019-12-23 12:58:04 +05:00

383 lines
15 KiB
Python
Executable file

# 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 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.common.network import create_dev, delete_network_interface, find_free_port
from ucloud.host import logger
from ucloud.shared import shared
from ucloud.settings import settings
from . import qmp
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 capture_all_exception(func):
@wraps(func)
def wrapper(*args, **kwargs):
try:
func(*args, **kwargs)
except Exception:
logger.exception('Unhandled exception occur in %s. For more details see Syslog.', __name__)
return wrapper
class VMM:
def __init__(self):
self.etcd_client = shared.etcd_client
self.storage_handler = shared.storage_handler
self.running_vms = []
def get_start_command_args(self, 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(
self.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)
for network_mac_and_tap in vm_networks:
network_name, mac, tap = network_mac_and_tap
_key = os.path.join(settings['etcd']['network_prefix'], vm_entry.owner, network_name)
network = self.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=settings['network']['vxlan_phy_dev'],
tap_id=tap,
ip=network_ipv6)
all_networks = self.etcd_client.get_prefix('/v1/network/', value_in_json=True)
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)
return command.split(" ")
def create_vm_object(self, vm_entry, migration=False, migration_port=None):
vnc_sock_file = tempfile.NamedTemporaryFile()
qemu_args = self.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)
@staticmethod
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)
@capture_all_exception
def create(self, vm_entry: VMEntry):
if self.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)
return None
else:
vm_hdd = int(bitmath.parse_string_unsafe(vm_entry.specs["os-ssd"]).to_MB())
if self.storage_handler.make_vm_image(src=vm_entry.image_uuid, dest=vm_entry.uuid):
if not self.storage_handler.resize_vm_image(path=vm_entry.uuid, size=vm_hdd):
vm_entry.status = VMStatus.error
else:
logger.info("New VM Created")
@capture_all_exception
def start(self, vm_entry: VMEntry, destination_host_key=None):
_vm = self.get_vm(self.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:
migration_port = find_free_port()
self.launch_vm(vm_entry, migration=True, migration_port=migration_port,
destination_host_key=destination_host_key)
else:
self.create(vm_entry)
self.launch_vm(vm_entry)
@capture_all_exception
def stop(self, vm_entry):
vm = self.get_vm(self.running_vms, vm_entry.key)
vm.handle.shutdown()
if not vm.handle.is_running():
vm_entry.add_log("Shutdown successfully")
vm_entry.declare_stopped()
shared.vm_pool.put(vm_entry)
self.running_vms.remove(vm)
delete_vm_network(vm_entry)
@capture_all_exception
def delete(self, vm_entry):
logger.info("Deleting VM | %s", vm_entry)
self.stop(vm_entry)
if self.storage_handler.is_vm_image_exists(vm_entry.uuid):
r_status = self.storage_handler.delete_vm_image(vm_entry.uuid)
if r_status:
shared.etcd_client.client.delete(vm_entry.key)
else:
shared.etcd_client.client.delete(vm_entry.key)
@capture_all_exception
def transfer(self, 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 = self.get_vm(self.running_vms, join_path(settings['etcd']['vm_prefix'], _uuid))
if vm:
tunnel = sshtunnel.SSHTunnelForwarder(
_host,
ssh_username=settings['ssh']['username'],
ssh_pkey=settings['ssh']['private_key_path'],
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 shared.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
self.running_vms.remove(vm)
vm.handle.shutdown()
source_vm.in_migration = False # VM transfer finished
finally:
tunnel.close()
@capture_all_exception
def launch_vm(self, vm_entry, migration=False, migration_port=None, destination_host_key=None):
logger.info("Starting %s" % vm_entry.key)
vm = self.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()
shared.vm_pool.put(vm_entry)
else:
vm_entry.vnc_socket = vm.vnc_socket_file.name
self.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=settings['etcd']['request_prefix']
)
shared.request_pool.put(r)
else:
# Typical launching of a vm
vm_entry.status = VMStatus.running
vm_entry.add_log("Started successfully")
shared.vm_pool.put(vm_entry)
@capture_all_exception
def maintenance(self, 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.
logger.debug("Starting Maintenance!!")
to_be_removed = []
for running_vm in self.running_vms:
with shared.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:
self.running_vms.remove(r)
# To check vm running according to etcd entries
alleged_running_vms = shared.vm_pool.by_status("RUNNING", shared.vm_pool.by_host(host.key))
for vm_entry in alleged_running_vms:
_vm = self.get_vm(self.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()
shared.vm_pool.put(vm_entry)
if _vm:
self.running_vms.remove(_vm)
def resolve_network(network_name, network_owner):
network = shared.etcd_client.get(join_path(settings['etcd']['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 = shared.vm_pool.by_owner(vm_entry.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], 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_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')
}
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'])