14:01:47 #startmeeting powervm_driver_meeting 14:01:48 Meeting started Tue Dec 13 14:01:47 2016 UTC and is due to finish in 60 minutes. The chair is esberglu. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:49 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:51 The meeting name has been set to 'powervm_driver_meeting' 14:02:08 o/ 14:02:11 No I didn't. I tried last night and couldn't get in. Now I can log in but everything I click on errors out 14:02:27 You need to request access to a project 14:04:17 That's a bizarre system... 14:04:27 Step 1: Sign in 14:04:33 Step 2: Request project access 14:04:38 Step 3: Wait indefinitely 14:04:53 I can't even request anything 14:05:01 hmm...pm him to see 14:05:03 When I click requests it says an error occured 14:05:08 #agile 14:05:11 we can just get started with the driver meeting for now 14:05:27 #topic openstack ci 14:05:43 404 more zuul mergers not found 14:06:01 Did you guys have a chance to look at the patch for that? 14:06:09 I did not. 14:06:33 Not yet, was busy in meetings 14:07:04 #action thorst and adreznec to do their reviews 14:07:14 Okay. I might be missing some stuff on it but I mostly just took the stuff we needed from that ansible zuul role 14:08:12 adreznec: Do you know how zuul knows that additional zuul mergers exist? 14:08:15 Ah there it is, I wasn't actually on the review. Looking now 14:08:26 I think they report into gearman 14:08:41 and register internally with zuul from there 14:11:13 Alright. The CI runs are still looking fine once they make it through the queue 14:11:22 But are queue is now backed up 77 jobs 14:11:27 *our 14:11:32 lol 14:11:37 so that's top CI priority 14:11:37 Awesome 14:11:39 Yeah 14:11:46 we need to get those extra nodes up 14:12:02 Jupiter froze my browser again 14:12:18 Man 14:12:20 What are you using 14:12:28 It worked fine for me in FF 14:12:32 IBMs Firefox 14:12:36 Lol 14:12:43 Yeah, I dumped the IBM FF a long time ago 14:12:53 Too many versions behind 14:13:28 Try chrome or safari or something 14:13:36 At least that'll tell you if it's the browser 14:13:56 #action esberglu Get into Jupiter cloud, make zuul mergers 14:14:17 I think that's pretty much it for openstack CI 14:14:30 Yeah, hard to do much until we solve that 14:15:12 #topic OSA CI 14:15:30 wangqwsh: Did you say we could use neo4 for CI? 14:15:57 no, i didnot 14:16:13 Okay. 14:16:33 thorst: Did you find out if neo50 could be used for OSA CI or if that was for SDN? 14:16:51 esberglu: I think that is for CI 14:16:58 we have two others waiting for Neutron SDN stuff 14:17:03 but I do not use noe4. If we need it for CI, we can use neo4 14:17:34 alright, so we have neo4 and 50 for OSA CI stuff :-) 14:17:39 Okay. I think it would be beneficial to have 2 systems for OSA CI since it is more resource heavy than openstack 14:17:53 And then keep neo14 for the openstack CI staging env. 14:18:11 #action: esberglu: Get neo4 and neo50 configured for CI 14:18:46 wangqwsh: Anything you need to do/get on neo4 before I start using it? 14:19:09 no, you can use it directly 14:19:13 Cool 14:19:20 I think a system rebuild would be good 14:19:27 it has traditional VIOSes and what not 14:19:41 Yeah. I think neo50 got installed with 2 VIOSes as well 14:19:55 So I will rebuild both 14:20:16 wait...maybe we need it that way 14:20:20 cause SSPs... 14:20:33 All of the other CI systems have 1 VIOS 14:20:35 I think we want single VIOS (for SSP) and then the SDN config on the NL 14:20:44 right...ok 14:20:48 I think we're saying the same thing 14:21:06 Yep 14:21:26 #topic Drivers 14:21:50 My status there is that I pushed up four WIP reviews after getting asked by mriedem where we were at 14:22:08 this is for the powervm integration into nova proper 14:22:18 we are behind, and need to make up ground there...which is a challenge with the holidays 14:22:33 my reviews did not have UT on them. wangqwsh has been doing an excellent job adding it 14:22:36 and we now need to review 14:22:46 https://review.openstack.org/#/c/391288/ 14:23:38 #action: all: review the WIP nova patches 14:24:45 The ocata 2 milestone is this week so I will be tagging the out of tree drivers for that 14:25:03 #action esberglu: Tag out of tree drives for ocata 2 milestone 14:28:03 Anything else for the drivers? 14:28:04 wangqwsh: for the nova driver review...as comments come in from people generally new to the powervm driver (ex. Ed Leafe)...they'll have great comments that we need to update the patch for 14:28:14 but we also want to make sure you then propose a corresponding patch back to nova-powervm's driver 14:28:27 to keep the out-of-tree driver in line with the in-tree driver 14:28:30 does that make sense? 14:29:22 ok... I do not sync comments to nova-powervm 14:29:29 I will do it later 14:29:46 sure 14:29:48 thx! 14:29:54 :) 14:31:00 #action wangqwsh to push updates back into nova-powervm based on review comments on nova WIP patches 14:31:05 that's all I had for the driver bits 14:31:14 #topic General Discussion 14:31:34 So these additional zuul mergers should work around the problem for now 14:31:54 But it still comes down to the poor network performance interacting with git.o.o 14:32:05 yeah, but that'll take a while 14:32:07 That kind of sounds like a combination issue 14:32:09 At one point, 1 zuul merger was handling WAY more volume than we are seeing now 14:32:13 Between git.o.o and the POK lab network 14:32:18 we need to get a ticket open to the network team 14:32:26 and give them endpoints they can work with 14:32:35 they may need to escalate to the back bone provider... 14:34:55 Can one of you guys chase that at some point? I don't think I know enough about the network 14:35:30 pm me a source IP and we'll get a ticket open...as well as recreate steps 14:35:34 I'll open it to the network team 14:35:52 but, if they end up having to work with the backbone provider...expect months 14:35:58 also, I think this is just a solid idea anyway 14:37:06 anything else? 14:37:15 Yeah 14:37:32 wangqwsh: Did you see my note about CI deployments? 14:37:38 yes 14:38:01 I said in there to try deploying the staging environment. Hold off on that for now while we work through this zuul merger think 14:38:07 thing 14:38:09 but i am working on WIP patch sets. so I will check it later 14:38:35 ok 14:38:54 Okay. I will let you know when it is not being used an you can try to deploy it then 14:39:05 sure, thanks 14:39:56 The last thing I had. If you guys want a quick look at how the CI is performing on nova you can look here 14:40:01 http://ci-watch.tintri.com/project?project=nova 14:40:09 And scroll down to IBM PowerVM CI 14:40:26 hah! we're so behind! 14:40:30 but excellent to be in there 14:41:07 I need to figure out why sometime the CI reports as IBM PowerVM CI and sometimes as IBM PowerVM CI check 14:41:41 That covers everything from my end 14:41:54 Anyone else have final topics? 14:42:07 thorst: what is the deadline of WIP-x patch sets? 14:42:13 this week? 14:42:54 wangqwsh: Lets stabilize the first one this week. Get a good head start on second and try to have the stable mid next week. 14:43:29 I also need to make a WIP 5 for the VIF bits. 14:43:56 ok, got it 14:44:25 many reviewers will be out in the next couple of week...starting on Friday, at least in the US, many people will be off till Jan 2nd 14:44:40 but we want as much of this lined up for review as possible on Jan 2. 14:45:48 ok 14:49:06 #endmeeting