[Openstack-operators] [Nova] Significance of Error Vs Failed status

David Medberry openstack at medberry.net
Wed May 11 22:05:08 UTC 2016


So I'm a big ol' -0- don't care on this. We've never used that list before
(but will now). Seems like it would be useful though to have it the same
for l-m and cold migration.

On Wed, May 11, 2016 at 9:27 AM, Andrew Laski <andrew at lascii.com> wrote:

>
>
>
> On Wed, May 11, 2016, at 11:10 AM, David Medberry wrote:
>
> Kekane,
>
> Hi,
>
> This setting, how does it display in the "nova show $UUID" or in the
> "openstack server show $UUID"? Ie, I don't want a VM showing ERROR state if
> the VM itself is not in error. A failed migration doesn't leave the VM down
> (well, not always) but error generally implies it is down. If this is more
> of an internal status, then +1. I'll look at the code shortly but wanted to
> get a reply off first.
>
>
> To clarify, this is only about the state of a migration not an instance.
> If as an admin you list or show your migrations this would affect how that
> is displayed. Nothing about the instance, or how it's displayed, will
> change.
>
>
>
>
> ALSO: It would have been very very helpful to see "live-migration" in the
> subject line.
>
>
> -d
>
> On Wed, May 11, 2016 at 12:55 AM, Kekane, Abhishek <
> Abhishek.Kekane at nttdata.com> wrote:
>
> Hi Operators,
>
>
>
> Could you please provide your opinion on below mail. I need to discuss
> this in coming nova meeting (12 May, 2016).
>
>
>
> Thank you,
>
>
>
> Abhishek Kekane
>
>
>
> *From:* Kekane, Abhishek [mailto:Abhishek.Kekane at nttdata.com]
> *Sent:* Monday, May 09, 2016 7:22 AM
> *To:* openstack-operators at lists.openstack.org
> *Subject:* [Openstack-operators] [Nova] Significance of Error Vs Failed
> status
>
>
>
> Hi All,
>
> In Liberty release, we had upstream [1] a security fix to cleanup orphaned
> instance files from compute nodes for resize operation. To fix this
> security issue, a new periodic task '_cleanup_incomplete_migrations’ was
> introduced that runs on each compute node which queries for deleted
> instances and migration status in “error” status. If there are any such
> instances, then it simply cleanup instance files on that particular compute
> node.
>
> Similar issue is reported in LP bug [2] for Live-migration operation and
> we would like to use the same periodic task to fix this problem. But in
> case of live migration, the migration status is set to “failed” instead of
> “error” status if migration fails for any reason. This change was
> introduced in patch [3] when migration object support was added for live
> migration. Due to this inconsistency, the periodic task will not pickup
> instances to cleanup orphaned instance files. To fix this problem, we
> simply want to set the migration status to “error” in patch [4] same as
> done for resize operation to bring consistency to the code.
>
> We have discussed about this issue in the nova meeting [5] and decided
> that to the client, migration status 'error' vs. 'failed' should be
> considered the same thing, it's a failure. From operators point of view, is
> there any significance of setting migration status to 'error' or 'failed',
> if yes what is it and what impact it will have if migration status is
> changed from 'failed' to 'error'. Please provide your opinions on the same.
>
>
>
> [1] https://review.openstack.org/#/c/219299
>
> [2] : https://bugs.launchpad.net/nova/+bug/1470420
>
> [3] https://review.openstack.org/#/c/183331
>
> [4] https://review.openstack.org/#/c/215483
>
> [5]
> http://eavesdrop.openstack.org/irclogs/%23openstack-meeting/%23openstack-meeting.2016-05-05.log.html#t2016-05-05T14:40:51
>
> Thank You,
>
> Abhishek
>
>
>
> ______________________________________________________________________
> Disclaimer: This email and any attachments are sent in strictest confidence
> for the sole use of the addressee and may contain legally privileged,
> confidential, and proprietary data. If you are not the intended recipient,
> please advise the sender by replying promptly to this email and then delete
> and destroy this email and any attachments without any further use, copying
> or forwarding.
>
>
>
> ______________________________________________________________________
> Disclaimer: This email and any attachments are sent in strictest confidence
> for the sole use of the addressee and may contain legally privileged,
> confidential, and proprietary data. If you are not the intended recipient,
> please advise the sender by replying promptly to this email and then delete
> and destroy this email and any attachments without any further use, copying
> or forwarding.
>
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
>
>
> *_______________________________________________*
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
>
>
>
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20160511/ac906a5a/attachment.html>


More information about the OpenStack-operators mailing list