uncloud/ucloud/scheduler/main.py

94 lines
4.5 KiB
Python
Raw Normal View History

# 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)
2019-12-03 11:49:10 +00:00
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()