Thursday, 2017-04-20

*** chas has joined #openstack-powervm00:22
*** edmondsw has joined #openstack-powervm00:23
*** chas has quit IRC00:27
*** edmondsw has quit IRC00:28
*** thorst has joined #openstack-powervm00:37
*** thorst has quit IRC00:42
*** chas has joined #openstack-powervm00:43
*** chas has quit IRC00:47
*** chas has joined #openstack-powervm01:04
*** chas has quit IRC01:08
*** thorst has joined #openstack-powervm01:13
*** chas has joined #openstack-powervm01:25
*** thorst has quit IRC01:25
*** chas has quit IRC01:29
*** thorst has joined #openstack-powervm01:33
*** thorst has quit IRC01:36
*** chas has joined #openstack-powervm01:45
*** chas has quit IRC01:50
*** kjw3 has joined #openstack-powervm02:04
*** chas has joined #openstack-powervm02:06
*** thorst has joined #openstack-powervm02:07
*** chas has quit IRC02:11
*** thorst has quit IRC02:23
*** chas has joined #openstack-powervm02:27
*** chas has quit IRC02:32
*** chas has joined #openstack-powervm02:48
*** chas has quit IRC02:53
*** thorst has joined #openstack-powervm02:54
*** thorst has quit IRC02:54
*** jay1_ has joined #openstack-powervm02:58
*** chas has joined #openstack-powervm03:09
*** jay1_ has quit IRC03:12
*** chas has quit IRC03:14
*** thorst has joined #openstack-powervm03:26
*** k0da has quit IRC03:34
*** thorst has quit IRC03:44
*** chas has joined #openstack-powervm03:50
*** chas has quit IRC03:56
*** edmondsw has joined #openstack-powervm04:00
*** edmondsw has quit IRC04:04
*** shyama has joined #openstack-powervm04:07
*** chas has joined #openstack-powervm04:12
*** chas has quit IRC04:16
*** chas has joined #openstack-powervm04:33
*** chas has quit IRC04:37
*** thorst has joined #openstack-powervm04:41
*** thorst has quit IRC04:46
*** chas has joined #openstack-powervm04:54
*** chas has quit IRC04:58
*** chas has joined #openstack-powervm05:35
*** chas has quit IRC05:40
*** thorst has joined #openstack-powervm05:42
*** thorst has quit IRC05:46
*** chas has joined #openstack-powervm05:56
*** edmondsw has joined #openstack-powervm06:00
*** chas has quit IRC06:01
*** edmondsw has quit IRC06:05
*** jay1_ has joined #openstack-powervm06:06
*** chas has joined #openstack-powervm06:17
*** chas has quit IRC06:21
*** chas has joined #openstack-powervm06:38
*** chas has quit IRC06:42
*** chas has joined #openstack-powervm07:25
*** chas has quit IRC07:26
*** chas has joined #openstack-powervm07:26
*** thorst has joined #openstack-powervm07:44
*** thorst has quit IRC08:03
*** k0da has joined #openstack-powervm08:30
*** thorst has joined #openstack-powervm10:00
*** thorst has quit IRC10:04
*** k0da has quit IRC10:30
*** thorst has joined #openstack-powervm11:01
*** thorst has quit IRC11:06
openstackgerritArun Mani proposed openstack/nova-powervm master: Deploy of VM occasionally fails with OSError
*** shyama has quit IRC11:07
*** smatzek has joined #openstack-powervm11:23
*** smatzek has quit IRC11:25
*** smatzek has joined #openstack-powervm11:25
*** thorst has joined #openstack-powervm11:26
*** shyama has joined #openstack-powervm11:27
*** chas_ has joined #openstack-powervm11:31
*** openstackgerrit has quit IRC11:32
*** chas has quit IRC11:35
*** a1fisher has quit IRC12:11
*** edmondsw has joined #openstack-powervm12:12
*** jay1_ has quit IRC12:18
*** mdrabe has joined #openstack-powervm12:34
efriedadreznec yt?12:35
efriedor thorst ?12:36
thorstwhat's up12:41
efriedI've never been clear on exactly how requirements are really supposed to work.12:41
efriedLike, when it's called for and/or acceptable to bump g-r vs. u-c.12:41
thorstadreznec is the better one12:42
thorstor edmondsw maybe....but I think adreznec knows that best12:42
efriedI don't feel like it would actually *hurt* us to stay at 1.1.1 for the time being.12:42
efried1.1.2 doesn't have any actual function or bugfix material that we need.12:42
efriedSo we could probably wait until 1.1.3 to actually bump.12:42
efriedBut we don't really want to miss it in pike, and have to keep patching local2remote for another release worth.12:43
efriedThat said, the current CI setup is pretty unobtrusive.12:43
efriedAnd we don't need to keep the old local2remote patch around.12:43
efriedSo maybe we don't care at all.12:43
thorstwell, I think we can bump upper-requirements13:01
thorstthat's kinda the reason for that13:01
thorstwe don't necessarily keep those two in sync all the time...13:01
thorstthen that would allow CI to bump up13:01
thorstat least for master13:02
*** apearson has joined #openstack-powervm13:05
edmondswefried I agree with thorst... sounds like you'd want to bump u-c but not g-r13:12
efriedthorst edmondsw Okay, done.13:12
efriedThanks guys.13:12
adreznecefried: thorst Sorry, just got in. Yep, agreed on the u-c only bump for now since you're really trying to move up the upper bound of allowed releases and not remove the older releases from support13:19
*** apearson has quit IRC13:30
*** apearson has joined #openstack-powervm13:32
*** k0da has joined #openstack-powervm13:36
*** kjw3 has quit IRC13:38
*** smatzek has quit IRC13:39
*** dwayne has joined #openstack-powervm13:49
*** smatzek has joined #openstack-powervm13:59
*** esberglu has joined #openstack-powervm14:05
*** tjakobs has joined #openstack-powervm14:22
thorstefried: dang dude...nice14:24
*** a1fisher has joined #openstack-powervm14:30
*** kjw3 has joined #openstack-powervm14:33
efriedthorst adreznec esberglu edmondsw - spawn/delete #1 got sdague +W!14:36
thorstand fast...14:36
adreznecWait no, I changed my mind, I don't to have the driver in tree anymore!14:37
edmondswsomeone shoot adreznec14:37
efriedThey will physically kill you at the summit.14:37
adrezneclol probably14:37
adreznecupgraded to definitely if mikal is there14:38
edmondswefried you saw my comments in
efriededmondsw Yuh, was just about to fix.  Good call.14:43
edmondswI didn't look at why jenkins failed14:44
efriededmondsw Done.  I'll take a quick look at the failure, but it's not likely ours.14:45
efriededmondsw yeah, a timeout in a test.14:46
efriededmondsw I believe addressed by
*** dwayne has quit IRC14:47
efriedwhich is now merged.14:48
*** chas_ has quit IRC15:04
*** chas has joined #openstack-powervm15:04
*** chas has quit IRC15:09
thorstefried: I -1'd 45770715:32
efriedthorst ack15:34
*** chas has joined #openstack-powervm15:55
*** chas has quit IRC15:59
*** mdrabe has quit IRC16:00
*** mdrabe has joined #openstack-powervm16:07
*** dwayne has joined #openstack-powervm16:12
*** apearson has quit IRC16:25
*** apearson has joined #openstack-powervm16:25
edmondswefried took me too long to go through so you'll see my comments there instead of your newer patch set16:42
edmondswsome of them you've probably addressed, but I doubt all16:42
efriededmondsw ack16:42
*** esberglu has quit IRC16:55
*** esberglu has joined #openstack-powervm17:34
*** shyama has quit IRC17:53
*** shyama has joined #openstack-powervm17:54
thorstefried: I think tomorrow or so we should work with Jay to get devstack stood up with the storwise cinder some iscsi testing there18:22
thorstsince nbante's test isn't going super hot...and we need to validate LPM with iSCSI18:22
*** chas has joined #openstack-powervm18:25
*** nbante has joined #openstack-powervm18:26
nbanteI am trying to rerun playbook to configure external stoarge but setup-hosts.yml is failing with some setuptools dependecy issue.18:29
nbanteI am getting below error when I manually tried18:29
nbantepython /opt/ --isolated setuptools==34.2.018:29
*** chas has quit IRC18:30
nbanteRequirement already up-to-date: setuptools==34.2.0 in /usr/local/lib/python2.7/dist-packages18:30
nbanteRequirement already up-to-date: pip in /usr/local/lib/python2.7/dist-packages18:30
nbanteRequirement already up-to-date: packaging>=16.8 in /usr/local/lib/python2.7/dist-packages (from setuptools==34.2.0)18:30
nbanteCollecting appdirs>=1.4.0 (from setuptools==34.2.0)18:30
nbante  Using cached appdirs-1.4.3-py2.py3-none-any.whl18:30
nbanteRequirement already up-to-date: six>=1.6.0 in /usr/lib/python2.7/dist-packages (from setuptools==34.2.0)18:30
nbanteCollecting pyparsing (from packaging>=16.8->setuptools==34.2.0)18:30
nbante  Using cached pyparsing-2.2.0-py2.py3-none-any.whl18:30
nbanteInstalling collected packages: appdirs, pyparsing18:30
nbante  Found existing installation: appdirs 1.4.018:30
nbante    Uninstalling appdirs-1.4.0:18:30
nbante      Successfully uninstalled appdirs-1.4.018:30
nbante  Rolling back uninstall of appdirs18:30
nbanteany clue to fix this issue18:31
edmondswnbante I'm not sure you pasted enough there... what was the actual error?18:39
esbergluefried: Looks like the --keep-redundant-commits isn't working properly. All of the in-tree patches are failing to apply PS1 since it merged18:43
esbergluI can remove PS1 from the patch list18:43
efriedWell wtf18:44
esbergluBut that means any patches being tested that haven't rebased to that point will fail18:44
esbergluBut it's weird because it would have been seeing it as a redundant commit before it merged right?18:44
efriedesberglu Pretty sure gerrit automatically rebases if a dependency has merged.18:44
efriedMaybe that's the problem.  Hold on, thinking...18:45
esbergluNo I'm saying just random nova patches18:45
efriedOh, probably so.18:45
esbergluBut nbd if we post some failures from that18:45
esbergluAt least in my opinion18:46
efriedesberglu What currently-whitelisted tests rely on changes >118:46
efriedSayin, can we just remove everything from the patch list?18:46
efriedMebbe bust the whitelist back down to just change-1 status.18:46
efriedNow that things are moving.18:46
efriedAnd lockstep it up as we merge more changes.18:46
esbergluefried: There are a few tests that rely on later changes18:48
esbergluKeeping the whitelist in sync with the IT driver is gonna be a headache18:49
esbergluStarting to think a blacklist might be better18:49
esbergluWe can use the OOT blacklist as a starting point18:50
*** jpasqualetto has quit IRC18:50
esbergluAnd any further tests that we disable we can leave a comment about why18:50
esbergluSo we should be able to enable certain tests as functionality comes in18:50
nbanteedmondsw I am getting below exception18:51
nbanteTraceback (most recent call last):18:51
nbante  File "/tmp/tmpMcX3v4/", line 215, in main18:51
nbante    status =, args)18:51
nbante  File "/tmp/tmpMcX3v4/", line 342, in run18:51
*** nbante has quit IRC18:51
*** nbante has joined #openstack-powervm18:51
*** newbie|3 has joined #openstack-powervm18:51
edmondswnbante you'd better put that in or something18:53
edmondswefried esberglu I was wondering if this would happen... I think you have to mark the commits as dependent on another commit, not as redundant18:54
efriedesberglu edmondsw Also why I was a bit concerned with cherry-picking each individually, rather than review -d of the top one and rebase onto HEAD18:55
edmondswefried esberglu I thought there was an easier way, but here's at least one way:
efriedRight, that's the way I always do it when I'm working on a series.18:58
*** newbie|3 has quit IRC18:59
*** nbante has quit IRC18:59
*** nbante has joined #openstack-powervm19:00
edmondswefried esberglu
efriededmondsw Yup, almost exactly the way I do it.  Except I use rebase instead of review -x19:02
*** chas has joined #openstack-powervm19:02
efriedPossible review -x is nicer because maybe it automatically skips redundant intervening rebase steps.19:03
esbergluWe could probably use -d19:03
nbanteI pulled latest ansible code and now dependency error gone. Facing another issue, trying to debug.19:04
esbergluJust would need to reorder so that it happens before we pull in the actual patch being tested I think19:04
*** chas has quit IRC19:06
*** jpasqualetto has joined #openstack-powervm19:07
esbergluefried: Removing the first patch from the list on production for now so that we can get our other IT patches passing19:09
efriedesberglu k19:09
esbergluReworking the patching logic as well19:09
*** shyama has quit IRC19:10
*** nbante has quit IRC19:10
efriedthorst Thoughts on edmondsw's suggestion of using save_and_reraise_exception all over the place?19:24
thorstefried: where was that?  And it seems reasonable...I just want to make sure we flush it all back to OOT19:25
edmondswefried thorst this explains why:
efriedJust seems like unnecessary code/logic if we're doing something really simple in the except clause.  Like, are we really worried that LOG.* will muck with the exception context?19:27
edmondswefried it's not that... some other greenthread could19:27
thorstif it is a pattern that everyone else follows, then I'm good with it19:27
efriededmondsw *Any* greenthread switch??  No way.19:28
efriedThat'd have to be a python bug.19:28
edmondswthorst efried I grepped and see 168 other places in nova using this19:28
edmondswso yes, it's a pattern19:28
efriedWell, I grepped (stand by, will count) places that don't.19:29
edmondswyou can ask in the nova channel... I'd guess that's 56 bugs, but maybe they will tell you it's not necessary in some cases19:29
efriedSo it's like, one of those herringbone patterns where one color is used more than the other.19:30
efriedI don't object to changing 'em all, necessarily, just seems overly paranoid in most cases.19:30
edmondswask mriedem19:30
edmondswI'm good with whatever he says19:30
efriedWill do.  He and sdague didn't call it out in their reviews, but easy enough to miss I suppose.19:31
edmondswefried yeah, I figured they just overlooked it19:32
*** chas has joined #openstack-powervm19:38
*** jpasqualetto has quit IRC19:42
*** chas has quit IRC19:43
*** apearson has quit IRC19:50
edmondswefried so, from sdague's last comment, i would say save_and_reraise_exception isn't required to merge this, but if you have to spin a new patchset it should use it19:50
efriededmondsw Yeah, I'll put it in.  Just seems more complicated than it's worth for these little cases.19:51
edmondswI hear you about base python bug... I think it'd be more an eventlet issue, but that's probably what you meant19:51
efriedYeah, I suppose eventlet would be to blame in that case.19:51
efriedIt wouldn't be terribly hard to test.19:52
edmondswif you have the time :)19:52
efriedMake an except clause that deliberately switches greenthreads and raises some exception.19:52
efriedTime... and motivation.19:53
efriedI probably don't care enough to prioritize that.19:53
efriedI'd be more swayed by the possibility of LOG.* doing something freaky.19:53
*** jpasqualetto has joined #openstack-powervm19:54
efriededmondsw "it would be nice if there was a specific childclass of HttpError for 404s, and you could just expect that and not have to do this kind of check and reraise all over the place."19:56
edmondswefried yeah, if you had that the save_and_reraise thing would be a non-issue19:56
efriedThis is a good idea.  We'd have to put it in pypowervm and only use it in nova once it's landed in a version we can require.19:56
efriedYou feel like proposing the pypowervm change for it?19:57
edmondswefried yeah, I thought of that... that's the rub19:57
edmondswefried I'm tied up at the moment with higher priorities19:57
efriededmondsw We're pretty likely to need another requirements bump before RC1, so it's not out of the realm of possibility.19:57
edmondswefried is there a todo list we could throw that on?19:57
efriedBut will need to be done in a later change set.19:57
efriededmondsw Launchpad bug.19:58
edmondswefried sure, I can open a bug19:58
efriedCool man.
*** apearson has joined #openstack-powervm20:02
openstackLaunchpad bug 1685016 in pypowervm "HttpError is not granular enough" [Undecided,New]20:03
efriededmondsw Thanks.20:03
*** smatzek has quit IRC20:09
*** apearson has quit IRC20:14
*** chas has joined #openstack-powervm20:14
*** chas has quit IRC20:18
edmondswefried LOG.exception("PowerVM error during destroy.", instance=instance) ... what is the instance=instance for?20:29
efriededmondsw So glad you asked.20:29
efriedThe logger nova uses knows to look for an 'instance' kwarg and if found, includes the instance UUID in the log message template.20:30
efriedMakes debugging way easier.20:30
efriedWorks for any LOG.* in nova.20:30
efriededmondsw Aaargh, lost an edit in an aborted rebase.  Didn't change spelling of goofy consts in vm.py20:32
edmondswefried and didn't add that 2nd blank line after imports20:32
efriededmondsw That one too.20:33
efriedI started addressing the comments on the wrong change set.20:33
edmondswok, I'll stop looking and let you fix20:33
efriedTried to stash and restore, but got a conflict that was gonna be more trouble than worth to resolve, so started over.20:33
efriedBut missed those.20:33
efriedStand by for fix.20:33
efriededmondsw K, done.20:34
edmondswefried why are you setting both id and _uuid to the same thing, rather than one being instance.uuid and one being the output from get_pvm_uuid?20:44
efriededmondsw Probably not.  Looking...20:44
efriededmondsw Probably a bug from OOT too, in fact.20:45
edmondswadded review comment as a question20:45
edmondswI've gotta run20:46
efriedI don't see the id field being used anywhere, but you're right, it's probably intended to be the instance ID.20:47
*** edmondsw has quit IRC20:51
*** edmondsw has joined #openstack-powervm20:51
*** edmondsw has quit IRC20:56
*** thorst has quit IRC21:11
*** edmondsw has joined #openstack-powervm21:20
*** edmondsw has quit IRC21:25
*** chas has joined #openstack-powervm21:26
*** apearson has joined #openstack-powervm21:30
*** chas has quit IRC21:31
*** esberglu has quit IRC21:46
*** esberglu has joined #openstack-powervm21:46
*** jpasqualetto has quit IRC21:47
*** esberglu has quit IRC21:51
*** dwayne has quit IRC21:55
*** chas has joined #openstack-powervm22:02
*** jpasqualetto has joined #openstack-powervm22:04
*** thorst has joined #openstack-powervm22:06
*** mdrabe has quit IRC22:06
*** thorst has quit IRC22:07
*** chas has quit IRC22:07
*** apearson has quit IRC22:08
*** k0da has quit IRC22:22
*** chas has joined #openstack-powervm22:38
*** thorst has joined #openstack-powervm22:39
*** chas has quit IRC22:43
*** tjakobs has quit IRC22:54
*** thorst has quit IRC22:56
*** kjw3 has quit IRC23:00
*** chas has joined #openstack-powervm23:14
*** chas has quit IRC23:19
*** chas has joined #openstack-powervm23:50
*** thorst has joined #openstack-powervm23:53
*** chas has quit IRC23:55

Generated by 2.14.0 by Marius Gedminas - find it at!