yahoo-eng-team team mailing list archive

[Alexey Stupnikov <1994983@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

Steps to reproduce:

1. Start an instance on a compute node (src compute).
2. Destroy src compute. Wait till nova-compute server will go down. 
3. Run 'openstack server stop' for an instance.
4. Run 'server set --state error' to the instance to be able to evacuate the instance (no longer needed if fix for bug #1978983 presents)
5. Evacuate the instance to another compute node (dst compute).
6. Start src compute.
7. Confirm that the evacuated instance is stopped after src compute come online.

This behavior is a bug because src compute could come up after few
months and shutdown some important instance. It looks like this behavior
is caused by cast RPC call for SRC compute to stop an instance which
sits in the queue until compute is back online and then causes DB state
change.

RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=2130112

** Affects: nova
     Importance: Undecided
     Assignee: Alexey Stupnikov (astupnikov)
         Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1994983

Title:
  Instance is stopped on destination node after evacuation if stop call
  was issued before evacuation

Status in OpenStack Compute (nova):
  New

Bug description:
  Steps to reproduce:

  1. Start an instance on a compute node (src compute).
  2. Destroy src compute. Wait till nova-compute server will go down. 
  3. Run 'openstack server stop' for an instance.
  4. Run 'server set --state error' to the instance to be able to evacuate the instance (no longer needed if fix for bug #1978983 presents)
  5. Evacuate the instance to another compute node (dst compute).
  6. Start src compute.
  7. Confirm that the evacuated instance is stopped after src compute come online.

  This behavior is a bug because src compute could come up after few
  months and shutdown some important instance. It looks like this
  behavior is caused by cast RPC call for SRC compute to stop an
  instance which sits in the queue until compute is back online and then
  causes DB state change.

  RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=2130112

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1994983/+subscriptions