Tuesday, 2018-01-09

*** edmondsw has joined #openstack-powervm00:36
*** esberglu has quit IRC00:37
*** openstackgerrit has joined #openstack-powervm02:07
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer-powervm master: Updated from global requirements  https://review.openstack.org/53195702:07
openstackgerritOpenStack Proposal Bot proposed openstack/networking-powervm master: Updated from global requirements  https://review.openstack.org/53196602:19
openstackgerritOpenStack Proposal Bot proposed openstack/nova-powervm master: Updated from global requirements  https://review.openstack.org/53196802:23
*** edmondsw has quit IRC02:30
*** adreznec has quit IRC03:07
*** adreznec has joined #openstack-powervm03:08
*** adreznec has quit IRC03:16
*** adreznec has joined #openstack-powervm03:21
*** esberglu has joined #openstack-powervm04:16
*** esberglu has quit IRC04:17
*** chhavi has joined #openstack-powervm04:37
*** AlexeyAbashkin has joined #openstack-powervm07:56
*** chhavi__ has joined #openstack-powervm08:29
*** chhavi has quit IRC08:29
*** chhavi__ has quit IRC09:48
*** chhavi__ has joined #openstack-powervm09:49
*** chhavi__ has quit IRC10:57
*** AlexeyAbashkin has quit IRC11:02
*** AlexeyAbashkin has joined #openstack-powervm11:02
*** apearson__ has quit IRC13:11
*** edmondsw has joined #openstack-powervm13:21
openstackgerritMerged openstack/ceilometer-powervm master: Updated from global requirements  https://review.openstack.org/53195713:40
openstackgerritMerged openstack/networking-powervm master: Updated from global requirements  https://review.openstack.org/53196613:44
*** esberglu has joined #openstack-powervm13:48
*** apearson__ has joined #openstack-powervm13:49
*** AlexeyAbashkin has quit IRC13:52
openstackgerritMerged openstack/nova-powervm master: Updated from global requirements  https://review.openstack.org/53196813:52
*** gman-tx has joined #openstack-powervm13:57
*** chhavi__ has joined #openstack-powervm13:58
*** AlexeyAbashkin has joined #openstack-powervm14:00
esberglu#startmeeting powervm_driver_meeting14:00
openstackMeeting started Tue Jan  9 14:00:24 2018 UTC and is due to finish in 60 minutes.  The chair is esberglu. Information about MeetBot at http://wiki.debian.org/MeetBot.14:00
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
*** openstack changes topic to " (Meeting topic: powervm_driver_meeting)"14:00
openstackThe meeting name has been set to 'powervm_driver_meeting'14:00
esberglu#topic In-tree Driver14:02
*** openstack changes topic to "In-tree Driver (Meeting topic: powervm_driver_meeting)"14:02
esbergluThe pypowervm uc change merged yesterday14:02
esbergluOVS has passed CI and is ready for reviews14:02
esbergluI ran SEA CI IT manually and it looked good and merged last night14:03
edmondswI pinged mriedem and sdague to start reviewing yesterday afternoon14:03
efriedI merged the *-powervm requirements bumps this morning.14:03
esbergluHowever it's broken on the actual SEA patch, but works for everything else14:03
edmondswdo we know why?14:04
esbergluWhen it tries to apply the OVS & SEA patches onto the SEA patch it gets a merge failure14:04
esbergluWe hit this last release and I thought we fixed it14:04
edmondswdid we only fix it OOT?14:04
esbergluBut we've changed a few things with how we set up the repos since then14:04
esbergluNo this would probably hit that as well, just not a very common case14:05
esbergluThat's first on my list today14:05
edmondswI will let mriedem and sdague know the OVS is fully ready, SEA we're still working a CI issue14:06
esbergluedmondsw: Cool tx14:06
esbergluThe vSCSI patch is looking good functionally. I need to clean up the comment throughout a couple of the files that are outdated, missing params, returns, etc.14:06
esbergluThe UT is passing tox locally, but some of the tests are failing the gerrit tox14:07
efriedHave you -r'd locally?14:07
esbergluefried: Not recently, but I have at some point since they started failing upstream14:09
esbergluIt's various tests in this file, not the same ones every time14:09
efriedhm, that doesn't bode well.14:09
esbergluI think that they may just be taking too long?14:10
esbergluThese tests call the top level attach/detach_volume methods and test the entire flows in each test14:11
edmondswdoes sound like a potential timeout issue14:13
esbergluWhich is not how most of the other UT for the IT work has been done, that was mostly smaller sets of code in each UT14:13
edmondswyeah, UT should be smaller14:13
esbergluI based them off of the OOT tests, I've been trying to keep them as in sync as I can14:13
esbergluThink it's worth reworking?14:13
edmondswwe might have to14:13
esbergluIf that is indeed the issue14:13
edmondswyou could try bumping the timeouts first and see if that resolves them14:14
edmondswto confirm the issue is what we think it is14:14
esbergluI can't bump the timeouts, they're passing locally and I can't control the upstream zuul14:14
edmondswoh I meant locally14:15
esbergluIt's just one file, it shouldn't be that much work to break the UT into smaller sections14:15
edmondswk, let's start there14:16
efriedor reduce the timeouts locally and see if the same kind of errors pop up.14:17
esbergluSure I can try that14:17
esbergluThat's all I had for IT, any other thoughts from you guys?14:17
edmondswefried will you be able to review the SEA and OVS patches today?14:17
efriedJust finished the OVS patch.14:18
efriedWill look at the SEA patch.14:18
efriedBeen ignoring vSCSI because WIP and Zuul -114:18
edmondswesberglu let us know when we should start looking at vSCSI, or if there are parts of it we should start looking at, like the non-test bits?14:19
esbergluI would appreciate reviews for vSCSI sooner rather than later14:19
esbergluThe UT is ready for review except for test_vscsi.py14:20
edmondswok, I'll review that today14:20
esberglu#topic Out-of-tree Driver14:22
*** openstack changes topic to "Out-of-tree Driver (Meeting topic: powervm_driver_meeting)"14:22
efriedI merged the *-powervm requirements bumps this morning.14:22
esbergluThe only thing I had OOT was the SaveBDM stuff from yesterday14:23
edmondswthat was OOT?14:24
edmondswI thought that was part of the IT vSCSI work14:24
esbergluWell I'm assuming the same errors I was hitting IT will hit OOT, but don't have an OOT stack right now to confirm14:24
edmondswoh, sure14:24
esbergluThe plan from yesterday was to confirm that it hits OOT as well, and remove if it's not needed there either14:25
esbergluI'm not sure if it's needed for live migration or something14:26
edmondswyep, absolutely14:26
edmondswall changes that we've made in the IT patches should be worked back to OOT14:26
edmondswfor OVS and SEA as well as vSCSI14:26
edmondswI'll fix the tests for that bdm cleanup patch I threw up yesterday... should have that up again in a few min14:27
edmondswI'm not thinking of anything else OOT of the top of my head14:28
esbergluI'll put together a list on the etherpad14:28
esberglu#topic Device Passthrough14:29
*** openstack changes topic to "Device Passthrough (Meeting topic: powervm_driver_meeting)"14:29
esbergluefried: Anything to discuss here?14:30
efriedI can summarize the current status of the work in nova if you like, but in general we've just got quite a lot of work to squeeze in before feature freeze.  I'm heads down on that.14:31
* efried apparently has multiple heads14:32
*** tjakobs has joined #openstack-powervm14:32
edmondswjust don't start talking to yourself :)14:32
edmondswwe can probably skip the summary and give you time to do the work for today14:33
efriedWhereas I wouldn't object to reviews from you guys, that's not strictly necessary; and there's really nothing else y'all can do to help without significant ramp-up.14:33
edmondswesberglu there was also a summary on the ML a couple days ago14:33
efriedThough edmondsw and charles may want to review the report client stuff out of interest since you're going to be working up similar code for pvc14:33
edmondswefried link?14:33
efriedSeries starting at https://review.openstack.org/#/c/521685/14:34
efriedGoal this week will be to clear some of those WIPs at the top14:35
efried(and get those bottom ones merged - that's on cores now)14:35
efriedlooks like the bottom two are in the gate, cool.14:36
efriedNothing else from me.14:36
efried[looks like the bottom two are in the gate, cool.] <== so edmondsw don't -1 them!14:36
edmondswlol... sure14:36
esbergluAlright sounds like we can move on14:37
esberglu#topic PowerVM CI14:37
*** openstack changes topic to "PowerVM CI (Meeting topic: powervm_driver_meeting)"14:37
esbergluNot much to talk about here other than the issue already mentioned14:38
edmondswping me when you figure that one out?14:38
esbergluedmondsw: ok14:38
edmondswI don't have anything else14:39
esbergluAfter feature freeze I'd like to be heads down on CI for a bit14:39
openstackgerritMatthew Edmonds proposed openstack/nova-powervm master: cleanup private bdm methods  https://review.openstack.org/53187414:39
esbergluI want to upgrade the undercloud to pike, update all of the management services14:39
esbergluAnd have a ton of other small changes/cleanup to do14:40
edmondswok, tx for the heads up... I think we can afford that after ff14:40
*** gman-tx has quit IRC14:41
esbergluOh and get depends-on working14:41
efriedmeh, low priority14:41
esberglu#topic Open Discussion14:41
*** openstack changes topic to "Open Discussion (Meeting topic: powervm_driver_meeting)"14:41
esbergluNothing else from me14:42
*** openstack changes topic to "This channel is for PowerVM-related development and discussion. For general OpenStack support, please use #openstack."14:44
openstackMeeting ended Tue Jan  9 14:44:19 2018 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:44
openstackMinutes:        http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2018/powervm_driver_meeting.2018-01-09-14.00.html14:44
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2018/powervm_driver_meeting.2018-01-09-14.00.txt14:44
openstackLog:            http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2018/powervm_driver_meeting.2018-01-09-14.00.log.html14:44
*** gman-tx has joined #openstack-powervm15:10
esbergluedmondsw: efried: I see what the problem is with the patching, not sure the best way to fix it15:33
esbergluRight now we merge the change being tested into whichever branch is being tested15:33
esbergluThen when we try to apply the patches found in patching.conf, we search the git log for the commit and skip it if it is found15:34
esbergluThe problem is that when we merge the initial change into master, the commit number from the changes being tested aren't in the git log15:34
esbergluonly a commit number from the merge of those changes into the branch15:34
esbergluSo you have whatever the last commit number is on master, and then a commit number for the merge15:34
esbergluWe weren't hitting this last release because we weren't merging the changes into the branch15:35
esbergluJust checking out whatever change we were testing15:36
edmondswesberglu are you saying the same thing is being patched twice?15:36
esbergluSo the commit chain (in this case OVS, then SEA) would be found in the git log and each patch would be skipped15:36
edmondswonce as the commit we're trying to test, and again because it's one of the reviews in our patch list?15:36
esbergluedmondsw: Yes15:37
edmondswesberglu so can't you then simply add a check to our patch list processing to ignore a patch listed there if it's the one we're trying to test?15:37
efriedyou wouldn't know15:40
esbergluedmondsw: That would work if it's a single commit. But not for commit chains. For example if we're testing SEA we don't want to apply the OVS patch15:40
esbergluThat's why we were checking the git log15:40
efriedAt some point we were doing it by changenumber,patchnumber, weren't we?15:40
edmondswthat "yeah" was to esberglu15:41
efriedthough I guess that doesn't help either.15:41
efriedI would say we could do the check-if-in-the-chain based on the Change-Id15:41
efriedBut then we wouldn't be sure to get the correct commit level.15:41
efried(to be clear, if we had the right change at the wrong level, we should fail - resolving that automatically is a challenge that *might* be solvable, but not worth it.)15:42
esbergluCould we merge the other way? Merge master into the change being tested instead of the change being tested into master?15:43
esbergluThat might retain the commit numbers in the git log15:43
edmondswI don't think that would work15:43
efriedya know, the git log *does* have all the commit hashes, if you look at it as a tree.  Not sure how to automate that, but it ought to be possible.15:44
efriedesberglu Do you have a repo downloaded somewhere that's in this condition?  (Merge commits but not original commits)15:44
esbergluefried: One sec, I can get one back into that state15:45
efriedesberglu I think if you add --all to your git log command, you'll pick up the missing commits.15:46
esbergluefried: Cool thanks that works!15:54
esbergluI can't believe this was working at all tbh15:56
esbergluThat {{ nova_chain_base_commit }} is a remnant from when this was in the ansible playbooks15:56
esbergluIt must just not be ignoring it I guess15:57
esbergluefried: Do you remember why we had the { nova_chain_base_commit }}..HEAD in that git log command? It just limits how far back the git log goes?16:04
efriedYeah.  What would the alternative be?16:05
esbergluefried: Do we actually need that? Just trying to decide if I should replace it somehow or rip it out16:05
esbergluI don't really want to be constantly updating the base commit there16:06
efriedWell, if you don't do that, you'll get the entire git history from the beginning of time.  I doubt you want that.16:06
efriedNo, we should be figuring that out dynamically.16:06
efriedIt should be e.g. origin/master16:06
efriedmdrabe edmondsw esberglu Take a look at mdbooth's latest post to openstack-dev ML (block device features).  We should see if we can make use of that.16:10
edmondswefried ack16:11
esbergluefried: Wouldn't that also go back to the beginning of time?16:12
efriedesberglu origin/master..HEAD will get you only the commits between origin/master and HEAD, inclusive.16:13
*** tjakobs has quit IRC16:13
efriedHm, the --all seems to blow it up.16:14
efriedWell, I can seem to use .. and --graph together to get the right set of commits.  But then you'd have to parse out the ascii art, which is doable but sucky.16:16
efriedThere must be a way.  RingTFM...16:16
efriedesberglu try --no-merges16:20
efried(instead of --all)16:22
esbergluefried: Worked16:22
esbergluefried: Oh cool that only prints the two commits in the chain16:24
esbergluovs and sea I mean16:25
esbergluefried: Thanks for the help. Change is up 623716:27
esbergluI'll do a full run through of the script quick on both SEA and OVS changes16:28
esbergluedmondsw: efried: One more quick change to that patch16:36
efriedesberglu What does that do?16:36
esbergluefried: Prevents a text editor from opening when doing the pull16:36
efriedwhy would that happen?  If the pull creates a merge commit?16:37
efried...which I guess maybe it sometimes would, which is why you were having this problem with merge commits in the first place?16:37
openstackgerritMerged openstack/nova-powervm master: cleanup private bdm methods  https://review.openstack.org/53187416:38
esbergluefried: Yeah16:38
efriedesberglu Questions for ya in -nova16:43
efriededmondsw maybe you know the answer offhand16:45
*** AlexeyAbashkin has quit IRC17:02
*** chhavi__ has quit IRC17:14
edmondswesberglu I added some comments on vSCSI17:34
edmondswstill need to look at vscsi.py and the tests, but I have to run17:35
edmondswwill be offline this afternoon17:35
esbergluedmondsw: Okay tx. Will take a look after lunch17:36
*** edmondsw has quit IRC17:58
*** edmondsw has joined #openstack-powervm18:01
*** edmondsw has quit IRC18:06
*** AlexeyAbashkin has joined #openstack-powervm18:26
*** AlexeyAbashkin has quit IRC18:30
*** esberglu has quit IRC18:31
*** esberglu has joined #openstack-powervm18:31
*** apearson__ has quit IRC18:57
*** apearson__ has joined #openstack-powervm19:02
*** apearson__ has quit IRC19:28
*** apearson__ has joined #openstack-powervm20:22
*** AlexeyAbashkin has joined #openstack-powervm22:00
*** AlexeyAbashkin has quit IRC22:05
*** gman-tx has quit IRC22:16
*** zerick_ has quit IRC22:33
*** zerick has joined #openstack-powervm22:34
*** zerick has quit IRC22:49
*** apearson__ has quit IRC22:59
*** chhavi__ has joined #openstack-powervm23:11
*** chhavi__ has quit IRC23:15
*** AlexeyAbashkin has joined #openstack-powervm23:22
*** AlexeyAbashkin has quit IRC23:27

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!