Tuesday, 2018-01-16

*** k0da has joined #openstack-powervm00:08
*** k0da has quit IRC00:28
*** csky has quit IRC00:54
*** csky has joined #openstack-powervm01:25
*** csky has quit IRC02:24
*** csky has joined #openstack-powervm02:26
*** edmondsw has quit IRC02:34
*** csky has quit IRC02:43
*** csky has joined #openstack-powervm03:09
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer-powervm master: Updated from global requirements  https://review.openstack.org/53386003:18
*** csky has quit IRC03:34
*** edmondsw has joined #openstack-powervm03:47
*** edmondsw has quit IRC03:52
*** csky has joined #openstack-powervm04:06
openstackgerritOpenStack Proposal Bot proposed openstack/networking-powervm master: Updated from global requirements  https://review.openstack.org/53390604:17
openstackgerritOpenStack Proposal Bot proposed openstack/nova-powervm master: Updated from global requirements  https://review.openstack.org/53391204:20
*** chhavi has joined #openstack-powervm04:31
*** chhagarw has joined #openstack-powervm04:31
*** csky has quit IRC05:39
*** csky has joined #openstack-powervm05:46
*** csky has quit IRC06:00
*** csky has joined #openstack-powervm06:27
*** csky has quit IRC07:03
*** edmondsw has joined #openstack-powervm07:24
*** edmondsw has quit IRC07:28
*** csky has joined #openstack-powervm07:29
*** csky has quit IRC08:08
*** AlexeyAbashkin has joined #openstack-powervm08:15
*** csky has joined #openstack-powervm08:35
*** csky has quit IRC08:59
*** edmondsw has joined #openstack-powervm09:12
*** k0da has joined #openstack-powervm09:14
*** edmondsw has quit IRC09:17
*** csky has joined #openstack-powervm09:26
*** csky has quit IRC09:51
*** csky has joined #openstack-powervm10:18
*** AlexeyAbashkin has quit IRC10:23
*** AlexeyAbashkin has joined #openstack-powervm10:26
*** chhavi has quit IRC10:28
*** chhagarw has quit IRC10:29
*** csky has quit IRC10:38
*** edmondsw has joined #openstack-powervm11:00
*** csky has joined #openstack-powervm11:04
*** edmondsw has quit IRC11:05
*** csky has quit IRC11:30
*** edmondsw has joined #openstack-powervm12:03
*** csky has joined #openstack-powervm12:48
*** csky has quit IRC13:21
*** esberglu has joined #openstack-powervm13:59
edmondswesberglu mtg?14:01
esberglu#startmeeting powervm_driver_meeting14:01
openstackMeeting started Tue Jan 16 14:01:19 2018 UTC and is due to finish in 60 minutes.  The chair is esberglu. Information about MeetBot at http://wiki.debian.org/MeetBot.14:01
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:01
*** openstack changes topic to " (Meeting topic: powervm_driver_meeting)"14:01
openstackThe meeting name has been set to 'powervm_driver_meeting'14:01
*** chhagarw has joined #openstack-powervm14:01
*** chhavi has joined #openstack-powervm14:01
esberglu#topic In-tree Driver14:02
*** openstack changes topic to "In-tree Driver (Meeting topic: powervm_driver_meeting)"14:02
esbergluJust waiting for cores and updating as reviews come in14:02
edmondswwe have until Friday to get everything in, as I understand it14:02
*** csky has joined #openstack-powervm14:03
edmondswefried should we ping folks again?14:03
efriedI... guess?14:03
efriedDid gibi re-review the OVS patch?14:04
edmondswno. I wasn't sure if you'd already pinged him again14:04
edmondswI can ping folks, I just don't want to ping folks you've already pinged and be a nuisance14:04
efriedI hit up gibi just now14:05
edmondswand stephenfin, I see. I'll let you ping folks since you're already doing that. Thanks14:06
efriedNote that mriedem is out this week (and maybe next)14:07
edmondswyeah, I saw14:07
edmondswalright, we'll see how it goes14:08
esbergluedmondsw: Are you sure it has to be this week? I thought this week was the freeze for stuff like oslo, os-vif, etc.14:09
efriedHaving patches "ready" is going to make it more likely they'll still be allowed to merge after the actual freeze date.14:09
edmondswpretty sure it's this week... next week is clients14:09
edmondswbut... there is some hope if they slip this week, as efried said14:09
efriedJan 25th is ff: https://wiki.openstack.org/wiki/Nova/Queens_Release_Schedule14:09
efriedThis Thurs is non-client lib14:10
efriedNote that that's the openstack schedule.  Nova usually follows it to within a few days, but sometimes not exactly.14:10
edmondswJan 25 is the ff for clients, Jan 18 (yeah, Thurs, not Fri) is the ff for everything else14:10
efriedwhoops, sorry14:10
efriedthat *is* the nova schedule.14:10
edmondswoh, one more thing IT... we have esberglu's patch for the feature docs adding powervm and mine to add taskflow as a requirement14:12
edmondswI just bumped the latter to keep up with g-r, so lost your +1s there14:12
efriedlink pls14:13
efriedsave me finding it.14:13
edmondswI think that's all IT?14:13
efriedoh, the g-r taskflow bump merged?14:13
esberglu#topic Out-of-tree Driver14:13
*** openstack changes topic to "Out-of-tree Driver (Meeting topic: powervm_driver_meeting)"14:13
edmondswefried yep14:13
esbergluI've been working through some random backlog items14:14
efriedI strenuously object on moral grounds to the two-spaces-to-one thing.14:14
esbergluefried: Doesn't matter to me, I just wanted to see if I could get the bash right14:15
edmondswI was gonna say we probably didn't NEED to do that... I like consistency, but it's not effort we need to spend14:15
edmondswI assume he wrote a bash script to make the changes14:16
efriedoh, gotcha.14:16
edmondswand for what it's worth, that seems to have worked14:16
edmondswso efried if you object to merging that, just -1 or -2 and we can abandon it. I won't push it. I was just going to say merge it if you didn't object, since the work is done and consistency14:16
efriedI was being (somewhat) facetious.14:17
efriedI do object to people getting uppity about one space, because they know not why they protest.14:17
edmondswlol... I hope I'm not getting uppity about it :)14:18
efriedAnd the reason doesn't apply to fixed-width.14:18
efriedI'll review the patch later.14:18
edmondswyeah, not a priority14:19
efriedThough tbh, I'm not sure it's worth muddying the git blame for five hundred lines.14:19
efriedBut I'm not going to approve it without at least glancing at it.14:19
efriedOkay, we've already spent too much time on it; let's move on.14:19
edmondswI just saw there's a patch up for refactoring volume drivers14:20
edmondswI haven't looked yet14:20
edmondswwill try to look at that today14:20
efriedon my list too.14:21
edmondswefried I wanted to ask what your plans are for 612114:21
edmondswI gave some comments there a while back14:22
efriedYeah.  It hasn't been a priority.14:22
edmondswprobably just sitting until we get past FF I assume14:22
efriedDoing what you request will be nontrivial and risky.14:22
efriedSo if I'm going to do it, I know I'm going to need a decent time box.14:22
edmondswdo you not think it needs doing?14:23
efriedIf I didn't think there was merit to the idea, I would have just started arguing right away :)14:23
efriedI'm still not sure it's *worth* doing, but empirically you're right.14:23
openstackgerritMerged openstack/ceilometer-powervm master: Updated from global requirements  https://review.openstack.org/53386014:23
edmondswbwarrum threw up a change for marking the encryption stuff expiremental, and I +1'd after one change14:24
edmondswI didn't +2 because I wanted to make sure you like it14:24
efriedYes, I definitely need to look at that.14:24
efriedMy plan was actually to do a diff against HEAD~2 - i.e. look at the encryption stuff as if it had been put in that way.14:25
efriedto make sure nothing was missed.14:25
efriedbut again, need a time box for that.14:25
efriedI think I'll probably get it this morning, because I'm *way* ahead of Jay in terms of code-for-reviews.14:25
edmondswI think that's all I had for OOT14:26
esberglu#topic Device Passthrough14:27
*** openstack changes topic to "Device Passthrough (Meeting topic: powervm_driver_meeting)"14:27
esbergluSame status as last week I'm assuming14:28
efriedsamo samo14:28
efriedThough things are looking slippy14:28
efriedbecause Jay's been out.14:28
edmondswefried I saw him online yesterday I think? Is he back 100%?14:29
efriedHe said 75%.  But yeah, he's back.14:29
edmondswflu sounds nasty this year14:29
efriedyeah, and I didn't get a flu shot.  I'm afeared.14:30
* edmondsw crosses fingers14:30
edmondswI did have a thought on this14:31
edmondswtriggered by the feature doc update that esberglu put up yesterday14:31
openstackgerritMerged openstack/networking-powervm master: Updated from global requirements  https://review.openstack.org/53390614:31
edmondswthere is another section to that doc for HPC, and it includes a GPU attach section14:31
edmondswwe should be working toward being able to include PowerVM as "complete" there14:32
edmondswso let's be thinking about that for Rocky14:32
edmondswall I had14:33
esberglu#topic PowerVM CI14:33
*** openstack changes topic to "PowerVM CI (Meeting topic: powervm_driver_meeting)"14:33
esbergluStarting to look into some stuff that I've been sitting on here14:34
openstackgerritMerged openstack/nova-powervm master: Updated from global requirements  https://review.openstack.org/53391214:34
esbergluThat's pretty up to date, I can go into more detail here if you want14:34
esbergluA lot of small cleanup14:35
esbergluThen the next big change will be the pike undercloud upgrade started in change 620014:35
edmondswesberglu will we be able to test OVS and SEA both in the same CI run?14:35
esbergluedmondsw: We should be able to get 2 pipelines going and run both on each patch14:36
esbergluBut OVS uses WAY more resources14:36
esbergluI stole my staging CI systems to test OVS and vSCSI. Once I get those back I can start messing with it14:37
edmondswI'll look over the TODO list in more detail post-meeting14:37
esberglu#topic Open Discussion14:39
*** openstack changes topic to "Open Discussion (Meeting topic: powervm_driver_meeting)"14:39
esbergluAnything else?14:39
efriedNothing from me14:40
edmondswprobably not for today14:40
*** openstack changes topic to "This channel is for PowerVM-related development and discussion. For general OpenStack support, please use #openstack."14:41
openstackMeeting ended Tue Jan 16 14:41:10 2018 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:41
openstackMinutes:        http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2018/powervm_driver_meeting.2018-01-16-14.01.html14:41
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2018/powervm_driver_meeting.2018-01-16-14.01.txt14:41
openstackLog:            http://eavesdrop.openstack.org/meetings/powervm_driver_meeting/2018/powervm_driver_meeting.2018-01-16-14.01.log.html14:41
edmondswesberglu efried can one of you +W on 6244?14:49
edmondswor I guess I could just do that14:49
efriedYeah, but done14:49
edmondswefried to answer your question... I don't think OpenStack was trying to get networkx 2.0 support in for Queens, so I don't know if we have to spin a release ASAP, but I'll check with prometheanfire14:51
efriedsounds good14:51
edmondswesberglu I added you to https://review.openstack.org/#/c/530816/15:22
edmondswwe had some conversations about this kind of thing when you were working on vSCSI for IT15:22
edmondswplease take a look and see how this aligns with those conversations15:22
*** apearson has joined #openstack-powervm15:23
edmondswand what you think15:23
esbergluedmondsw: ack15:27
*** esberglu has quit IRC15:32
*** esberglu has joined #openstack-powervm15:48
esbergluefried: edmondsw: Single test skip list update in 6246 when you get a chance16:06
esbergluedmondsw: OVS is +2 from gibi16:17
edmondswesberglu woot!16:18
*** chhagarw has quit IRC16:24
*** chhavi has quit IRC16:24
*** k0da has quit IRC16:50
*** AlexeyAbashkin has quit IRC17:11
*** openstackgerrit has quit IRC17:33
esbergluefried: edmondsw: Thoughts on https://review.openstack.org/#/c/422512/39/nova/virt/powervm/vif.py@105 ?17:44
edmondswno, I don't see how neutron could use this17:49
edmondswthey only care about networking. They don't care about VMs17:49
edmondswit's nova's job to tie VMs and networks together17:49
edmondswunless I'm missing something...17:49
efriedHow about a response that covers our bases17:50
efriedLike, "Pretty sure not, but if it did, we would want to pull it out in a future series"17:50
esbergluefried: What about the nova-net comment? Go along with gibi's comment and say we will remove all nova-net stuff and only claim neutron support in a future change?17:54
efriedsure, sounds good to me.17:54
edmondswI asked thorst if the OOT driver worked with nova-network and he said no18:01
edmondswsaid they tried making that work a while back, hence the comments18:21
esbergluedmondsw: Think I should go ahead and rip it out? I don't want to lose the reviews, but I left a comment saying it's not supported, so I suspect I will have to respin anyways18:36
esbergluOr just wait and see what they say18:36
edmondswyeah, I'm not sure. Maybe ask stephenfin on irc18:38
edmondswif he does want you to respin, he can suggest where we should say that only neutron is supported18:38
edmondswbut I'd rather just leave it if he's ok with that18:38
*** AlexeyAbashkin has joined #openstack-powervm18:43
efriedesberglu Know about git restack?18:45
esbergluefried: Nope18:45
efriedpip install git-restack18:45
efriedThen from the top of your series, run   git restack18:46
efriedChange 'pick' to 'edit' for the patch(es) you want to edit, save and quit the editor.18:46
efriedWhen done editing, git rebase --continue18:46
efriedThen git review will submit the whole series at once.18:46
*** AlexeyAbashkin has quit IRC18:48
efriedSo in this case if you're just editing the one patch, git restack will automatically do the rest of the rebasing for you.18:49
efriedIf it gets stuck, it'll stop and tell you so; you resolve conflicts just like you would any merge conflict and then git rebase --continue.18:49
edmondswthat's pretty cool18:50
edmondswesberglu if you're gonna respin, better pull out not only comments re: nova-net but also the code for it18:50
esbergluedmondsw: Yep18:50
edmondswand then better retest18:50
edmondswesberglu re: 6246... how is that update method run by that test? It doesn't look like it should be from what I've found18:54
edmondswdid you have a stacktrace? I'm curious now18:54
edmondswand wondering if rename is something that we need to add support for. I thought that would just work without any driver involvement (just rename in the nova db)18:55
edmondswI thought the references to rename in our driver were just for renaming things on the backend during resize/migration18:56
efriedWe rename the lpar, nah?18:56
edmondswdo we?18:56
esbergluedmondsw: Let me finish this and then I can dig up my notes from last week. I could be mistaken18:56
efriedBut yeah, I thought our LparBuilder crap took care of all that.18:56
efriedA little surprised if it doesn't.18:56
efriedor if we left any of that out.18:56
efriedI can't imagine we wanted to go through the pain of dissecting LparBuilder just so we could, what, reduce the LOC count or whatever?18:57
*** efried is now known as efried_nomnom18:57
edmondswefried_nomnom LparBuilder is in pypowervm, so same for IT and OOT18:58
efried_nomnomVMBuilder, then18:59
esbergluefried_nomnom: Man git-restack is smooth18:59
esbergluThanks for the tip18:59
edmondswefried_nomnom ah, yeah19:00
efried_nomnomesberglu For sure.  Still gotta be a little careful sometimes - mainly don't commit -a when you shouldn't - but it makes dealing with tall series WAY easier.19:00
efried_nomnomesberglu viz: https://review.openstack.org/#/c/533244/ -- 12 tall :)19:01
esbergluedmondsw: You're right about that test19:36
edmondswesberglu so do we need to revert that patch?19:41
edmondswis there another change we need instead?19:41
esbergluedmondsw: Yeah. And figure out what's actually happening19:41
edmondswefried ^19:41
esbergluedmondsw: ^19:43
edmondswesberglu you might want to figure it out before merging the revert19:46
edmondswjust to keep things working19:46
*** k0da has joined #openstack-powervm19:55
esbergluedmondsw: It seems like this should be failing every time. The server name gets updated to non-ascii value, which causes the delete to fail20:14
esbergluBut it always is the same non-ascii value20:15
esbergluComparing successful and failing but not seeing what is different yet20:15
edmondswesberglu where are the logs?20:17
esbergluYou can see the name gets changed to non-ascii on the success, with the terminate going through shortly after20:18
edmondswcould it be pulling different versions of taskflow and/or networkx ?20:21
edmondswbetween success and failure20:21
edmondswnot sure where to look for that here20:21
edmondswesberglu ^20:22
edmondswesberglu failure has taskflow 3.1.0 and success has 3.0.1, but same version of networkx (1.11)20:24
esbergluedmondsw: Where are you seeing that?20:24
edmondswesberglu because the error comes from networkx, called by taskflow:
esbergluedmondsw: I'm seeing 3.1.0 for both?20:27
edmondswesberglu oh, my bad20:28
edmondswyeah, 3.1.0 for both20:28
edmondswthere goes that theory20:28
esbergluedmondsw: efried_nomnom: OVS is ready for new round of +1s20:40
*** efried_nomnom is now known as efried20:42
edmondswesberglu gonna have a couple small comments20:43
edmondswstill looking20:43
edmondswesberglu ok done20:44
esbergluedmondsw: Done20:47
*** esberglu has quit IRC21:37
*** esberglu has joined #openstack-powervm21:50
*** chhagarw has joined #openstack-powervm22:20
*** chhavi has joined #openstack-powervm22:20
*** chhagarw has quit IRC22:24
*** chhavi has quit IRC22:24
edmondswesberglu did you see gibi's suggestion of adding a patch with a release note on powervm additions to the end of the series?22:49
edmondswI liked that idea...22:50
*** csky has quit IRC22:54
*** csky has joined #openstack-powervm22:55
*** edmondsw has quit IRC22:55
*** edmondsw has joined #openstack-powervm22:56
esbergluedmondsw: Yeah will do22:57
*** csky has quit IRC22:59
*** edmondsw has quit IRC23:00
*** csky has joined #openstack-powervm23:25

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