| *** esberglu has quit IRC | 00:01 | |
| *** tjakobs has quit IRC | 00:01 | |
| *** esberglu has joined #openstack-powervm | 00:29 | |
| *** esberglu has quit IRC | 00:33 | |
| *** esberglu has joined #openstack-powervm | 00:38 | |
| *** svenkat has joined #openstack-powervm | 01:13 | |
| *** jwcroppe has joined #openstack-powervm | 01:26 | |
| *** thorst has joined #openstack-powervm | 01:29 | |
| *** mdrabe has quit IRC | 01:34 | |
| *** jwcroppe has quit IRC | 01:41 | |
| *** jwcroppe has joined #openstack-powervm | 01:42 | |
| *** jwcroppe has quit IRC | 01:46 | |
| *** esberglu has quit IRC | 01:49 | |
| *** svenkat has quit IRC | 02:00 | |
| *** svenkat has joined #openstack-powervm | 02:03 | |
| *** esberglu has joined #openstack-powervm | 02:04 | |
| *** esberglu has quit IRC | 02:04 | |
| *** edmondsw has joined #openstack-powervm | 02:20 | |
| *** edmondsw has quit IRC | 02:24 | |
| *** svenkat has quit IRC | 02:36 | |
| *** edmondsw has joined #openstack-powervm | 02:52 | |
| *** esberglu has joined #openstack-powervm | 02:58 | |
| *** esberglu has quit IRC | 03:03 | |
| *** thorst has quit IRC | 03:04 | |
| *** thorst has joined #openstack-powervm | 03:05 | |
| *** esberglu has joined #openstack-powervm | 03:12 | |
| *** esberglu has quit IRC | 03:12 | |
| *** thorst has quit IRC | 03:21 | |
| *** edmondsw has quit IRC | 05:33 | |
| *** edmondsw has joined #openstack-powervm | 05:33 | |
| *** edmondsw has quit IRC | 05:38 | |
| *** tlian has quit IRC | 05:41 | |
| *** tjakobs has joined #openstack-powervm | 06:53 | |
| *** tjakobs has quit IRC | 06:59 | |
| *** openstackgerrit has quit IRC | 07:32 | |
| *** k0da has joined #openstack-powervm | 08:18 | |
| *** edmondsw has joined #openstack-powervm | 09:34 | |
| *** edmondsw has quit IRC | 09:39 | |
| *** k0da has quit IRC | 10:58 | |
| *** smatzek has joined #openstack-powervm | 12:00 | |
| *** smatzek_ has joined #openstack-powervm | 12:02 | |
| *** smatzek has quit IRC | 12:02 | |
| *** edmondsw has joined #openstack-powervm | 12:03 | |
| *** seroyer has joined #openstack-powervm | 12:34 | |
| *** jpasqualetto has joined #openstack-powervm | 12:37 | |
| *** edmondsw has quit IRC | 12:59 | |
| *** xia has joined #openstack-powervm | 13:01 | |
| *** tblakes has joined #openstack-powervm | 13:02 | |
| *** seroyer has quit IRC | 13:07 | |
| *** svenkat has joined #openstack-powervm | 13:08 | |
| *** seroyer has joined #openstack-powervm | 13:17 | |
| *** thorst has joined #openstack-powervm | 13:24 | |
| *** edmondsw has joined #openstack-powervm | 13:30 | |
| *** thorst has quit IRC | 13:33 | |
| *** thorst has joined #openstack-powervm | 13:34 | |
| *** edmondsw has quit IRC | 13:35 | |
| *** thorst has quit IRC | 13:36 | |
| *** edmondsw has joined #openstack-powervm | 13:39 | |
| *** smatzek_ has quit IRC | 13:40 | |
| *** thorst has joined #openstack-powervm | 13:40 | |
| *** thorst_ has joined #openstack-powervm | 13:44 | |
| *** edmondsw has quit IRC | 13:44 | |
| *** thorst has quit IRC | 13:45 | |
| *** tlian has joined #openstack-powervm | 13:46 | |
| *** esberglu has joined #openstack-powervm | 13:47 | |
| *** dwayne has quit IRC | 13:48 | |
| *** smatzek has joined #openstack-powervm | 13:49 | |
| *** tlian has quit IRC | 13:49 | |
| efried | esberglu Looks like we're not properly applying the pypowervm local2remote patch in-tree? | 13:53 |
|---|---|---|
| *** edmondsw has joined #openstack-powervm | 13:55 | |
| *** wangqwsh has joined #openstack-powervm | 13:59 | |
| esberglu | #startmeeting powervm_driver_meeting | 14:02 |
| openstack | Meeting started Tue Feb 21 14:02:12 2017 UTC and is due to finish in 60 minutes. The chair is esberglu. Information about MeetBot at http://wiki.debian.org/MeetBot. | 14:02 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 14:02 |
| openstack | The meeting name has been set to 'powervm_driver_meeting' | 14:02 |
| adreznec | o/ | 14:02 |
| esberglu | Looks like the CI is failing for in-tree and out-of-tree since last night | 14:03 |
| xia | o/ | 14:03 |
| efried | o/ | 14:04 |
| esberglu | But we can talk about it more in a topic | 14:04 |
| esberglu | #topic In Tree Driver | 14:04 |
| efried | esberglu I don't have this meeting on my calendar. | 14:05 |
| efried | Can you re-invite me? | 14:05 |
| *** mdrabe has joined #openstack-powervm | 14:05 | |
| efried | So I'm looking at the in-tree failure, and I believe it's because we're not applying the pypowervm patches correctly. | 14:06 |
| efried | The problem is here: | 14:06 |
| efried | git fetch "$PYPOWERVM_REPO" $commit && git cherry-pick FETCH_HEAD | 14:06 |
| efried | The fetch is failing - though I can't tell why - so the cherry-pick isn't happening. | 14:06 |
| efried | This should really be split out, so that the whole script fails if the fetch fails. | 14:07 |
| esberglu | Where are you seeing the fetch fail? I'm not seeing that | 14:07 |
| efried | it's subtle. | 14:07 |
| efried | http://184.172.12.213/80/427380/8/check/nova-in-tree-pvm/57d5ca6/console.html Look around 20:30:37 | 14:08 |
| efried | You see the git fetch, but you don't see the git cherry-pick | 14:09 |
| efried | Compare and contrast to 20:30:53-54 where the cherry-pick succeeds, producing output, and then the cherry-pick happens. | 14:09 |
| efried | This is the danger of using that && syntax. | 14:10 |
| efried | It's really only for when you don't care if the LHS fails. | 14:10 |
| efried | ...but the RHS is still dependent on the LHS. | 14:10 |
| efried | I can repro the failure locally. | 14:11 |
| efried | But it still isn't saying anything. | 14:11 |
| efried | No idea why it's not working. Will keep poking. | 14:11 |
| esberglu | I'm gonna turn off CI for now then. This is causing every run to fail | 14:12 |
| *** kylek3h has joined #openstack-powervm | 14:13 | |
| efried | It works when I use the corresponding refs/changes, but not when I use the commit hash. | 14:15 |
| esberglu | But the same format is working for the other patches? | 14:16 |
| *** thorst_ has quit IRC | 14:17 | |
| efried | Some of them yes, some of them no. | 14:17 |
| esberglu | I bet it | 14:17 |
| esberglu | is the ones that aren't merged yet | 14:17 |
| *** thorst has joined #openstack-powervm | 14:17 | |
| esberglu | Some of these patches are merged, some of them are not | 14:18 |
| efried | Well, that makes zero sense, but I can check it. | 14:18 |
| efried | Not that it matters. I'll put up a fix. | 14:18 |
| *** burgerk has joined #openstack-powervm | 14:22 | |
| esberglu | Idk. But the 2 not merged patches are the patches that fail. Looks like the unmerged nova patch also isn't working | 14:23 |
| esberglu | The nova patch fails with this message | 14:23 |
| esberglu | error: no such remote ref db98f3c1f409faf3d2b5ca1f8951240f0a11414e | 14:23 |
| esberglu | On the git fetch | 14:23 |
| esberglu | Either way. That's all I had for my side of things. Anyone else have discussion items? | 14:25 |
| *** smatzek has quit IRC | 14:31 | |
| *** seroyer has quit IRC | 14:36 | |
| esberglu | I will get the new invites out for this meeting and the thursday CI meeting. Thanks for joining | 14:38 |
| esberglu | #endmeeting | 14:39 |
| openstack | Meeting ended Tue Feb 21 14:39:02 2017 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 14:39 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2017/powervm_driver_meeting.2017-02-21-14.02.html | 14:39 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2017/powervm_driver_meeting.2017-02-21-14.02.txt | 14:39 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2017/powervm_driver_meeting.2017-02-21-14.02.log.html | 14:39 |
| efried | esberglu 4894 | 14:39 |
| *** kylek3h has quit IRC | 14:42 | |
| *** kylek3h has joined #openstack-powervm | 14:42 | |
| *** kylek3h has quit IRC | 14:43 | |
| *** kylek3h has joined #openstack-powervm | 14:43 | |
| *** kylek3h has quit IRC | 14:43 | |
| *** kylek3h has joined #openstack-powervm | 14:44 | |
| *** kylek3h has quit IRC | 14:44 | |
| *** kylek3h has joined #openstack-powervm | 14:45 | |
| *** kylek3h has quit IRC | 14:45 | |
| *** k0da has joined #openstack-powervm | 14:49 | |
| *** smatzek has joined #openstack-powervm | 14:54 | |
| *** kylek3h has joined #openstack-powervm | 15:08 | |
| *** wangqwsh has quit IRC | 15:10 | |
| *** burgerk has quit IRC | 15:16 | |
| *** thorst has quit IRC | 15:18 | |
| *** edmondsw has quit IRC | 15:18 | |
| *** seroyer has joined #openstack-powervm | 15:20 | |
| *** edmondsw has joined #openstack-powervm | 15:23 | |
| *** edmondsw has quit IRC | 15:26 | |
| *** edmondsw has joined #openstack-powervm | 15:26 | |
| *** edmondsw_ has joined #openstack-powervm | 15:29 | |
| *** edmonds__ has joined #openstack-powervm | 15:30 | |
| *** edmondsw has quit IRC | 15:31 | |
| *** thorst has joined #openstack-powervm | 15:31 | |
| *** edmonds__ has quit IRC | 15:31 | |
| *** edmondsw has joined #openstack-powervm | 15:31 | |
| *** thorst has quit IRC | 15:32 | |
| *** edmondsw_ has quit IRC | 15:33 | |
| *** edmondsw_ has joined #openstack-powervm | 15:33 | |
| *** edmondsw has quit IRC | 15:36 | |
| *** edmondsw has joined #openstack-powervm | 15:38 | |
| *** edmondsw_ has quit IRC | 15:38 | |
| *** edmondsw has quit IRC | 15:39 | |
| *** tjakobs has joined #openstack-powervm | 15:39 | |
| *** edmondsw has joined #openstack-powervm | 15:39 | |
| *** thorst has joined #openstack-powervm | 15:44 | |
| *** thorst has quit IRC | 15:44 | |
| *** thorst has joined #openstack-powervm | 15:44 | |
| *** xia has quit IRC | 15:45 | |
| *** burgerk has joined #openstack-powervm | 15:55 | |
| *** apearson_ has joined #openstack-powervm | 15:55 | |
| *** k0da has quit IRC | 15:58 | |
| thorst | efried: I was just pinged offline about https://review.openstack.org/#/c/435108/1 | 15:59 |
| thorst | It looks fine to me. I put a +2 | 15:59 |
| thorst | any concerns? | 15:59 |
| *** burgerk has quit IRC | 16:00 | |
| *** mdrabe has quit IRC | 16:03 | |
| *** burgerk has joined #openstack-powervm | 16:07 | |
| *** dwayne has joined #openstack-powervm | 16:10 | |
| *** mdrabe has joined #openstack-powervm | 16:10 | |
| *** tlian has joined #openstack-powervm | 16:12 | |
| *** thorst_ has joined #openstack-powervm | 16:16 | |
| *** thorst has quit IRC | 16:17 | |
| *** k0da has joined #openstack-powervm | 16:20 | |
| mdrabe | efried: I've got a migration failure happening in this call in nova-powervm | 16:20 |
| mdrabe | https://github.com/openstack/nova-powervm/blob/master/nova_powervm/virt/powervm/driver.py#L1535 | 16:21 |
| mdrabe | Problem is, migration comes back as a success even though that call fails | 16:21 |
| mdrabe | My question is, do you think it should fail out instead? | 16:21 |
| efried | mdrabe I'm looking. But we'll want thorst_ on that discussion. | 16:22 |
| efried | mdrabe How does that call fail? | 16:24 |
| efried | If it raises, the live_migration will fail. | 16:24 |
| mdrabe | efried, thorst_ yea nevermind, I'm looking at this wrong | 16:25 |
| mdrabe | Somehow the migration is still coming back as succeeded | 16:25 |
| efried | mdrabe How does that call fail? | 16:26 |
| mdrabe | efried: I'll pm you the stack trace | 16:26 |
| efried | mdrabe So what makes you think the migration is succeeding? | 16:30 |
| mdrabe | We get a message saying that it succeeded | 16:30 |
| mdrabe | And it actually shows as on the new host | 16:30 |
| mdrabe | It's weird, I need to keep looking. I jumped the gun thinking it was nova-powervm related with that call | 16:31 |
| *** kylek3h has quit IRC | 16:51 | |
| *** thorst_ has quit IRC | 17:00 | |
| *** thorst has joined #openstack-powervm | 17:05 | |
| *** edmondsw has quit IRC | 17:05 | |
| *** edmondsw has joined #openstack-powervm | 17:06 | |
| *** thorst has quit IRC | 17:07 | |
| *** thorst has joined #openstack-powervm | 17:07 | |
| *** edmondsw has quit IRC | 17:11 | |
| *** edmondsw has joined #openstack-powervm | 17:14 | |
| *** edmondsw_ has joined #openstack-powervm | 17:16 | |
| *** edmondsw has quit IRC | 17:18 | |
| *** edmondsw_ has quit IRC | 17:21 | |
| *** thorst has quit IRC | 17:23 | |
| *** thorst has joined #openstack-powervm | 17:28 | |
| *** thorst has quit IRC | 17:30 | |
| *** apearson_ has quit IRC | 17:32 | |
| *** thorst has joined #openstack-powervm | 17:32 | |
| *** kylek3h has joined #openstack-powervm | 17:38 | |
| *** nbante has joined #openstack-powervm | 17:38 | |
| *** kylek3h has quit IRC | 17:39 | |
| *** apearson has joined #openstack-powervm | 17:45 | |
| *** thorst has quit IRC | 17:46 | |
| *** thorst has joined #openstack-powervm | 17:51 | |
| *** edmondsw has joined #openstack-powervm | 17:52 | |
| *** thorst has quit IRC | 17:56 | |
| *** tjakobs has quit IRC | 17:56 | |
| *** thorst has joined #openstack-powervm | 17:56 | |
| *** edmondsw has quit IRC | 17:57 | |
| *** edmondsw has joined #openstack-powervm | 17:58 | |
| *** edmondsw_ has joined #openstack-powervm | 18:03 | |
| *** edmondsw has quit IRC | 18:03 | |
| *** edmondsw_ has quit IRC | 18:03 | |
| *** edmondsw has joined #openstack-powervm | 18:03 | |
| *** tjakobs has joined #openstack-powervm | 18:04 | |
| *** openstack has joined #openstack-powervm | 18:14 | |
| *** edmondsw has quit IRC | 18:15 | |
| *** nbante has quit IRC | 18:25 | |
| *** thorst_ has joined #openstack-powervm | 18:27 | |
| *** apearson has quit IRC | 18:28 | |
| *** thorst has quit IRC | 18:29 | |
| *** edmondsw has joined #openstack-powervm | 18:31 | |
| *** edmondsw has quit IRC | 18:41 | |
| *** edmondsw has joined #openstack-powervm | 18:42 | |
| *** tlian has quit IRC | 18:42 | |
| *** edmondsw has quit IRC | 18:43 | |
| *** edmondsw has joined #openstack-powervm | 18:43 | |
| *** edmondsw has quit IRC | 18:45 | |
| *** edmondsw has joined #openstack-powervm | 18:46 | |
| *** edmondsw has quit IRC | 18:48 | |
| *** edmondsw has joined #openstack-powervm | 18:48 | |
| *** smatzek has quit IRC | 18:50 | |
| *** smatzek has joined #openstack-powervm | 18:58 | |
| *** apearson has joined #openstack-powervm | 19:00 | |
| *** tlian has joined #openstack-powervm | 19:05 | |
| *** thorst_ has quit IRC | 19:05 | |
| *** thorst has joined #openstack-powervm | 19:09 | |
| *** jwcroppe has joined #openstack-powervm | 19:17 | |
| *** apearson has quit IRC | 19:17 | |
| *** apearson has joined #openstack-powervm | 19:20 | |
| *** tlian has quit IRC | 19:22 | |
| *** thorst has quit IRC | 19:32 | |
| *** jwcroppe has quit IRC | 19:35 | |
| *** jwcroppe has joined #openstack-powervm | 19:36 | |
| *** jwcroppe has quit IRC | 19:40 | |
| *** openstackgerrit has joined #openstack-powervm | 19:44 | |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: In-tree backports: Task UT, Delete args https://review.openstack.org/435114 | 19:44 |
| *** thorst has joined #openstack-powervm | 19:50 | |
| *** edmondsw has quit IRC | 19:51 | |
| *** edmondsw has joined #openstack-powervm | 19:52 | |
| efried | esberglu Aborted? What's going on there? | 19:55 |
| esberglu | Trying to figure it out. A few runs got stuck in a loop looking for an existing LU | 19:57 |
| *** edmondsw has quit IRC | 19:57 | |
| *** edmondsw has joined #openstack-powervm | 19:58 | |
| *** edmondsw has quit IRC | 20:04 | |
| *** jwcroppe has joined #openstack-powervm | 20:04 | |
| *** edmondsw has joined #openstack-powervm | 20:04 | |
| *** edmondsw has quit IRC | 20:07 | |
| *** edmondsw has joined #openstack-powervm | 20:07 | |
| *** thorst_ has joined #openstack-powervm | 20:21 | |
| *** thorst has quit IRC | 20:23 | |
| efried | esberglu adreznec thorst_ Booyah, first two change sets are golden. Just finished polishing the third (localdisk) for thorst_ review comments, ready for next review. | 20:34 |
| *** tlian has joined #openstack-powervm | 20:36 | |
| *** tlian has quit IRC | 20:41 | |
| *** tlian has joined #openstack-powervm | 20:44 | |
| *** jwcroppe has quit IRC | 20:54 | |
| esberglu | efried: We are going to need to update the in tree local.conf for localdisk to pass | 20:55 |
| esberglu | oslo_config.cfg.RequiredOptError: value required for option volume_group_name in group [powervm] | 20:55 |
| efried | Indeeeed. | 20:55 |
| efried | I just did that, too ;-) | 20:56 |
| thorst_ | do we have any volume groups on these systems | 20:56 |
| efried | I mean, added that as a required opt. | 20:56 |
| *** thorst_ is now known as thorst_afk | 20:56 | |
| efried | The SSP change set will make it not-required again. | 20:56 |
| efried | esberglu Doesn't actually matter if we make the opt not-required, since we'd be trying to use the localdisk driver anyway. | 21:01 |
| efried | So I think we need to figure out a way to run CI with localdisk driver enabled. | 21:02 |
| *** edmondsw has quit IRC | 21:03 | |
| *** edmondsw has joined #openstack-powervm | 21:04 | |
| thorst_afk | efried: can we do the SSP one first? | 21:04 |
| efried | thorst_afk Not easily ;-) | 21:05 |
| efried | But yeah, if we did that, we would never have to get localdisk working ;-) | 21:05 |
| thorst_afk | I think that may be easier than trying to get localdisk into the CI | 21:05 |
| efried | Anyone here think it's a bad idea for us to run the CI at least once with localdisk? | 21:05 |
| thorst_afk | me. | 21:06 |
| thorst_afk | because we would have to fundamentally retool the CI for it | 21:06 |
| thorst_afk | I think it makes sense to run tempest tests against a custom setup | 21:07 |
| thorst_afk | but not refactor the entire CI setup for it when we've tuned it fundamentally for SSPs | 21:07 |
| efried | Okay. I'll get on that. Going to be a nontrivial reswizzle, though. | 21:08 |
| efried | Like, more than just a rebase. | 21:08 |
| efried | Cause we geared localdisk as if it was going to stand alone, then SSP as if it was going on top. | 21:08 |
| efried | So gotta reverse that. | 21:08 |
| *** edmondsw has quit IRC | 21:11 | |
| *** burgerk has quit IRC | 21:23 | |
| *** smatzek has quit IRC | 21:26 | |
| thorst_afk | efried: yeah...sorry. I think that's probably the better approach | 21:27 |
| thorst_afk | we had originally done localdisk cause it made more sense initially...but...this is probably better | 21:27 |
| efried | thorst_afk First slapping console on top of power on/off/reboot... | 21:44 |
| efried | thorst_afk esberglu Done - that one should be able to pass CI now. | 21:46 |
| *** tjakobs has quit IRC | 21:46 | |
| efried | https://review.openstack.org/#/c/409402/ | 21:46 |
| efried | Think I'm going to start a new SSP change set, rather than trying to rebase wangqwsh's | 21:47 |
| *** tjakobs has joined #openstack-powervm | 21:49 | |
| *** thorst_afk has quit IRC | 21:51 | |
| *** tblakes has quit IRC | 21:58 | |
| *** svenkat has quit IRC | 22:02 | |
| *** thorst_afk has joined #openstack-powervm | 22:03 | |
| *** thorst_afk has quit IRC | 22:04 | |
| mdrabe | efried: about that post_method call in the live_migration driver method. I'm not sure if this is a defect, but if that method fails when it's called nothing is done about it | 22:08 |
| efried | Again, define "fails". | 22:09 |
| mdrabe | raises an exception | 22:09 |
| efried | Then the live_migration method will raise an exception. | 22:09 |
| mdrabe | Correct but the manager doesn't do anything about it | 22:09 |
| efried | But if the compute manager doesn't do anything about that, that's not under our control. | 22:09 |
| efried | So... is that a problem? | 22:09 |
| mdrabe | I was thinking the driver should do something about it, but wasn't sure | 22:09 |
| efried | Have you found the code where the compute manager is running the live_migration method? | 22:10 |
| efried | Is that guy handling specific exceptions specially? | 22:10 |
| mdrabe | So that post_method, in my case, and maybe all cases is _post_live_migration | 22:10 |
| mdrabe | It handles some exceptions | 22:10 |
| mdrabe | It's _post_live_migration defined in the compute manager | 22:10 |
| efried | Could we "fix" it by making our code raise an exception that's handled differently? | 22:10 |
| mdrabe | So the problem that's specifically happening in my case is | 22:11 |
| efried | We call the post_method within our version of live_migration. We could catch whatever exception you think ought to be fatal, and reraise something that the compute manager _does_ treat as fatal. | 22:11 |
| efried | If you think that's the right thing to do. | 22:11 |
| mdrabe | _post_live_migration fails in terminate_connection | 22:11 |
| mdrabe | This causes the exception to bubble up and not be handled | 22:12 |
| mdrabe | So there's no cleanup or post live migration handling on the destination node | 22:12 |
| mdrabe | This leaves the instance in the driver's migrating dictionary on the destination | 22:12 |
| mdrabe | And causes some issues for subsequent migrations | 22:13 |
| mdrabe | The thing is it's the driver that's supposed to handle the cleanup I think | 22:13 |
| mdrabe | The manager isn't going to do anything | 22:13 |
| efried | So does that mean our live_migration method needs a 'finally' that cleans that stuff up regardless of whether the _post fails? | 22:15 |
| *** tlian has quit IRC | 22:15 | |
| mdrabe | I was thinking about maybe just invoking that _migration_exception_util in the event the post_method fails | 22:16 |
| mdrabe | The _post_live_migration should really be handling some more exceptions | 22:18 |
| esberglu | efried: The console patch is passing CI. I'm setting up staging to see what additional tests we should add to the in-tree whitelist | 22:21 |
| efried | esberglu Cool beans. Remember we have to merge the first one before we update that. | 22:24 |
| efried | esberglu How about the power on/off/reboot one? | 22:24 |
| efried | Do that one first. | 22:24 |
| *** tblakes has joined #openstack-powervm | 22:24 | |
| esberglu | Yep that was the plan. | 22:25 |
| efried | mdrabe I'm not looking at the code, but I'm instinctively not liking the idea of using private methods. | 22:25 |
| mdrabe | efried: I'm boiling it down to a hole in nova | 22:26 |
| efried | mdrabe Bug? | 22:26 |
| efried | Sounds like it | 22:26 |
| mdrabe | Yea, it's a fairly rare case I think | 22:27 |
| efried | Though it may be tough to push through, given it may break legacy. | 22:27 |
| mdrabe | Tbh I don't even know what I'd want to happen | 22:27 |
| mdrabe | It's not like we'd reverse migrate the VM back to the source at that point | 22:27 |
| efried | Well, if it's "fail the migration" (and what, roll back?) then yeah, we have a bug. | 22:27 |
| efried | If it's "make sure we clean up anyway", then we can fix that in our driver with a `finally`. | 22:28 |
| mdrabe | The cleanup that'd need to happen is more or less everything that happens in here: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L5522-L5636 | 22:29 |
| mdrabe | libvirt doesn't do any driver side cleanup either | 22:32 |
| efried | oho. | 22:33 |
| efried | The thing that's failing is calling into the consumer's cinder driver code. | 22:33 |
| efried | Sooooo... maybe that thing shouldn't "fail". | 22:34 |
| mdrabe | good point | 22:34 |
| mdrabe | ours does | 22:34 |
| efried | Question is, does consumer's terminate_connection get called from other places too? | 22:35 |
| mdrabe | I think so | 22:36 |
| mdrabe | Probably in the detach volume and delete flows | 22:36 |
| efried | Might want to spy a look at other cinder drivers and see whether/how they avoid exceptions in their terminate_connection methods. | 22:37 |
| mdrabe | Yea, I know who to talk to for that too | 22:38 |
| efried | Think he's in Atlanta. | 22:39 |
| *** tlian has joined #openstack-powervm | 22:40 | |
| mdrabe | Yea... got a few folks on the storage side though | 22:40 |
| *** esberglu has quit IRC | 22:48 | |
| *** tlian has quit IRC | 22:49 | |
| *** tlian has joined #openstack-powervm | 22:59 | |
| *** dwayne has quit IRC | 23:02 | |
| *** seroyer has quit IRC | 23:10 | |
| *** k0da has quit IRC | 23:10 | |
| *** mdrabe_ has joined #openstack-powervm | 23:12 | |
| *** mdrabe has quit IRC | 23:16 | |
| *** mdrabe_ has quit IRC | 23:17 | |
| *** tlian has quit IRC | 23:18 | |
| *** esberglu has joined #openstack-powervm | 23:21 | |
| *** esberglu has quit IRC | 23:26 | |
| *** svenkat has joined #openstack-powervm | 23:29 | |
| *** apearson has quit IRC | 23:45 | |
| *** tjakobs has quit IRC | 23:51 | |
| *** tlian has joined #openstack-powervm | 23:59 | |
| *** tblakes has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!