16:20:12 #startmeeting cinder 16:20:13 Meeting started Wed Aug 27 16:20:12 2014 UTC and is due to finish in 60 minutes. The chair is jgriffith. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:20:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:20:16 The meeting name has been set to 'cinder' 16:20:18 YAYYY for fungi !!! 16:20:35 #topic J3 status 16:20:43 ok... now we're official again 16:20:54 \o/ 16:20:59 Note bot was toast so we're 20 minutes in without topic 16:21:13 flip214: is my understanding correct? 16:21:40 What should we do for driver cert till CI is working? Create a bug and submit cert results? 16:21:44 jgriffith: well, I'm not sure. with some luck DRBDmanage should have snapshots by September ... 16:21:56 Arkady_Kanevsky: we have to rally and fix issues :) 16:22:05 flip214: yeah, that's too late :( 16:22:06 but for integration/testing/debugging that's a bit late, I know. 16:22:19 flip214: it'll be item 1 in Kilo 16:22:31 let's talk about third party ci 16:22:36 any chance of a maintenance release at the end of year? 16:22:37 oh. 16:22:37 #topic third-party-ci status 16:23:02 flip214: not for drivers... but we push to master and people can pull that 16:23:11 flip214: we can talk in cinder channel after if you like? 16:23:25 flip214: we do maintenance releases for bugs 16:23:31 no new features or drivers though 16:23:42 ok... third party CI 16:23:50 jgriffith: it's okay, thanks. At least the uncertainty has vanished! 16:24:06 some folks have mentioned that things seem to be a bit ..... *unstable* as of late 16:24:24 particularly rigs that were passing are not passing any longer 16:24:26 couple of things: 16:24:28 looks like we've got some issue with tempest and/or devstack. not sure that all of them are bugs in sinder 16:24:44 Check tempest and devstack git history, see if you can figure out when/what went wrong 16:24:57 This tempest bug is still not merged: https://review.openstack.org/#/c/112535/ 16:25:41 about test_volume_boot_pattern 16:26:00 xyang1: is that believe to *fix* the issues? 16:26:12 jgriffith, one issue we ran into was that we were using the old openstack pypi mirrors, which didn't have the new requirements...hence lots of failures. 16:26:25 ok... let's all back up a second 16:26:41 jgriffith: that's supposed to fix some issues related to test_volume_boot_pattern that we run into 16:26:54 so, it seems some people are having issues with CI that aren't necessarily driver related 16:27:04 jgriffith: not sure if it will fix all issues 16:27:12 The most common ones I've heard are test_volume_boot_pattern 16:27:34 I've also personally noticed a lot of flakiness in attaching of volumes created with encryption types 16:27:42 https://review.openstack.org/#/c/116290/ - this fixes some issue with tempest and rbd backend (about 5 tests) 16:27:53 everything else seems to run *ok* and consistently 16:28:20 e0ne: that's Ceph specific 16:28:31 jgriffith: yep 16:28:48 e0ne: I'm really only concerned with issues people are having with running that *are not* related to their driver 16:28:57 sorry:( 16:28:59 or don't appear to be related to their driver 16:29:02 e0ne: no worries 16:29:24 jgriffith: Just testing running tempest following setup from Jay Pipes instructions fails. Nothing related to specific driver. 16:29:51 smcginnis: ok, so are you saying the actual setup of the systems? Or the tests? 16:30:15 jgriffith: Some issues with the setup. Seem to have resolved them, now it's just down to running the tests. 16:30:23 What I'd like to see people do if possible is log bugs with a "third-party-ci-fail" tag 16:30:26 jgriffith: Haven't even gotten to trying to test our driver. 16:30:31 how many CI systems are online and allowed to give feedback on non-sandbox cinder stuff? I only saw two on my review?? 16:30:49 but make absolutely certain it's not your driver, or your CI-Rig before doing so 16:31:16 rhe00: I think there's two vendors active right now HP and EMC 16:31:23 rhe00: IBM should be back online shortly 16:31:24 ok, that's what I saw 16:31:37 Who can help me with pylint errors? My patch is failing with "Module 'hashlib' has no 'sha256'". We may need to add this error to exception list 16:31:48 I'll add mine when I figure out the attach issues/failures I'm seeing 16:31:58 Murali: link to patch? 16:32:10 pylint 16:32:18 Murali: I'll take a look at it after the meeting and add comments if I can help 16:32:21 jgriffith: Hope to be. 16:32:36 Murali: yes... I understand pylint, which patch/review ID? 16:32:47 https://review.openstack.org/110068 16:32:50 so is the plan still to boot all drivers that don't have an active vendor CI, and if so, by when? 16:33:17 Murali: we currently have an exception like that for 'sha224'. it's likely lintstack.py needs to be modified to generate that exception as well 16:33:23 rhe00: let's focus on getting things up and running for now 16:33:30 ok 16:33:33 jgriffith:thanks John 16:33:41 rhe00: we can discuss things like driver removal etc later 16:33:50 Murali: you're very welcome 16:34:14 Ok... so here's a plan 16:34:30 DuncanT-: started a g-docs spreadsheet of third-party ci status 16:34:45 I propose we all update that with our current status 16:34:56 and start an issues page 16:35:06 jgriffith: I know I've seen it, but do you have a link? 16:35:10 where issues == tempest tests that fial for you 16:35:24 smcginnis: I don't, but I'll find it... was hoping DuncanT- would appear 16:35:43 jgriffith: +1 on that plan 16:36:09 Personally, my CI has a few tests failing (I agree on that plan, jgriffith), I still have to check whether it's config-related or not 16:36:24 smcginnis: https://docs.google.com/spreadsheets/d/1WCfGEm1XgK79vMMi67dxUcK_sw4FINQhzPbrQ_NRvxM/edit?usp=sharing 16:36:32 ^^ link to ci status sheet 16:36:43 jgriffith: Thank you much! 16:36:44 "You need permission" :( 16:37:02 yup, requires access. 16:37:09 jgriffith: could you, please, make it a bit more public? 16:37:11 Requested to DuncanT- 16:37:18 jgriffith: who supose to access that? 16:37:25 the one that i looked at and edited before didn't have access 16:37:29 jgriffith smcginnis https://docs.google.com/spreadsheets/d/1WCfGEm1XgK79vMMi67dxUcK_sw4FINQhzPbrQ_NRvxM/edit?usp=sharing_eil 16:37:30 restrictions* 16:37:53 because i can't access that either 16:38:13 and i thought it was an etherpad... 16:38:24 can't access it either 16:38:27 I thought there was an etherpad eariler 16:38:28 jgriffith: jungleboyj So we wait for DuncanT- ? 16:38:51 smcginnis: Yeah, I was supposed to have access but don't appear to. 16:38:56 I'll leave access permissions to DuncanT- 16:39:06 if he doesn't want to change it I'll start a new version 16:39:10 jgriffith: so the data from the etherpad was moved here? 16:39:14 should we all email DuncanT- with status and ask him to update it? 16:39:17 eharney: which etherpad :) 16:39:31 jgriffith: the one that driver owners were asked to fill out with information listing all the drivers. 16:39:39 which... i can't find now 16:39:40 https://etherpad.openstack.org/p/cinder-driver-maintainers ? 16:40:02 nahh... that's just the maintainers list 16:40:12 alright... 16:40:16 How's this 16:40:25 I'll create a new etherpad 16:40:51 unless somebody has an actual link to an existing one with this info? 16:41:38 * jungleboyj plays the crickets sound 16:41:49 https://etherpad.openstack.org/p/cinder-third-party-ci-status 16:42:37 jgriffith: where do we report issues? on this etherpad as well? 16:42:46 should we savage-edit ? :p 16:42:54 xyang1: yes, to start with 16:43:05 jgriffith: ok 16:43:10 xyang1: and of course if you think you've identified a valid bug it should be logged in LP 16:43:24 and use a tag 16:43:30 I'll pu tthat info on the etherpad 16:43:37 jgriffith: sure. I opened a bug for the boot pattern issue 16:44:06 xyang1: awesome 16:44:11 xyang1: there is an issue about the tesT_volume_boot_pattern ? Seems to be one I encounter. 16:44:17 Ok 16:44:42 joa: yes, patch is here: https://review.openstack.org/#/c/112535/ 16:44:55 Well, I have to go through my CI's log to put down which test modulesI encounter issues with 16:45:02 I suspsect configuration issues for some of them 16:45:14 should we put them all down, and then triage this later ? 16:45:15 joa: bug: https://bugs.launchpad.net/tempest/+bug/1351441 16:45:16 Launchpad bug 1351441 in tempest "addCleanup is missing from create snapshot test" [Undecided,Fix committed] 16:45:47 Ok, let's all see if we can't get some data on status on that etherpad (that now won't open for me) 16:46:00 joa: I'll add this info to etherpad as well 16:46:27 Gah, I was an hour out, sorry 16:46:28 alrighty... 16:46:35 #topic open-discussion 16:46:45 DuncanT-: yer forgiven, sir. 16:47:28 I'll try to make that spreadsheet fully public 16:47:37 DuncanT-: +2 16:47:38 so for those that haven't seen.... 16:47:50 https://docs.google.com/spreadsheets/d/1WCfGEm1XgK79vMMi67dxUcK_sw4FINQhzPbrQ_NRvxM/edit?usp=sharing 16:47:58 Hopefully now done 16:48:16 DuncanT-: right got access 16:48:17 thanks 16:48:19 DuncanT-: started a "kiss" version of setting up third-party-ci https://github.com/Funcan/kiss-ci 16:48:28 DuncanT-: I am in now. 16:48:41 I still need to push John's updates 16:48:43 I've taken some liberties on that, and I've also now got a version that uses ansible-playbook for everything 16:48:51 DuncanT-: How should we tell you to make some changes ? 16:48:59 the point being, use what you want, but if you're hung up on jenkins etc, don't be 16:48:59 I'll do that in the next few minutes before I get pulled off onto something else again 16:49:10 move on for now and get something running asap 16:49:12 joa: To teh spreadsheet? Just go ahead and edit 16:49:20 ookay, lets try :D 16:49:46 cant seem to be able to edit my companys status 16:49:52 isnt it write-locked ? 16:49:59 it looks read-only to me too 16:50:08 okay so i'm not alone 16:50:13 * joa felt lonely one second. 16:52:02 Well, I'll update Scality's status when it will be writeable :) 16:52:16 wait... hold up 16:52:24 let's decide here and now 16:52:28 one version of truth 16:52:35 +1 16:52:36 g-doc spreadsheet or etherpad 16:52:37 Could the "Build project" status URL be added into that spreadsheet ? 16:52:38 not both 16:52:50 +1, I find gdoc table more visual 16:52:57 I personally don't care which it is as long as it's publicly available 16:53:17 but I certainly don't want to have multiple versions floating around 16:53:20 i'd vote etherpad just because it's generally easier for people to pull up and edit (in my experience anyway) 16:53:32 +1 16:53:38 all in favor of etherpad show +1: 16:53:46 +1 16:53:46 +1 16:53:48 +1 16:53:50 +1 16:53:57 +1 16:53:58 +1 16:53:59 DuncanT-: Since you are the keeper of the CI's, any preference? 16:54:00 +1 16:54:09 I don't mind 16:54:20 +1 16:54:24 ok.... all in favor of a g-doc spreadsheet show +1: 16:54:28 +1 16:54:36 * joa feels lonely :D 16:54:52 +1 16:54:56 * jungleboyj can go either way 16:54:57 ok... that seems like a pretty clear win 16:54:59 jungleboyj: WTF 16:55:03 you can't vote for both 16:55:04 aahah 16:55:06 lol 16:55:06 etherpad +1 16:55:11 jgriffith: I felt bad for joa 16:55:15 haha 16:55:25 0 No score 16:55:26 etherpad is fine 16:55:28 in etherpad it's much easier to change other people's minds. 16:55:29 :) 16:55:35 jungleboyj: what a nice guy 16:55:41 Ok, let's try etherpad, but I'd like ot figure out if we can create a table in etherpad 16:55:42 kmartin: You know it! 16:55:47 and keep things somewhat organized 16:55:55 jgriffith: that was my.. internal worry ? 16:55:57 jgriffith: +1 16:56:04 if there can be a good table I'm all for it 16:56:11 +1 16:56:15 only gdoc's one is nice and visual 16:56:27 if we can have the same on etherpad, all the better 16:56:40 The spreadsheet is now editable by y'all in case that is useful 16:56:41 spreadsheet was prettier 16:56:55 oh boy... we've completely lost control :) 16:57:02 :) 16:57:05 wiki.openstack.org is also an option to consider 16:57:16 eharney: indeed, good point 16:57:25 eharney: good compromise 16:57:38 eharney: I'm making an exec decision... 16:57:41 eharney: I thought we were trying to make less options, not more! :) 16:57:42 wiki.openstack 16:57:49 I'll link it to cinder wiki page 16:57:54 done 16:57:58 it's easily editable but a little less unstable than an etherpad environment 16:58:01 sounds good to me... 16:58:11 also, logs the changes :)) 16:58:16 * DuncanT- also invites related reviews on https://review.openstack.org/116887 - I've got a few more to add but digging through emails is far harder than letting you all comment 16:58:30 just have to figure out concurrency so not all at once... :) 16:58:49 joa: etherpad actually does that, it's just got a tricky UI 16:58:53 * jungleboyj can't use the ASCII table generator I just found now. 16:59:02 eharney: oh. tricky it is, then 16:59:06 https://wiki.openstack.org/wiki/Cinder/third-party-ci-status 16:59:19 I'll get things flushed out (table etc) 16:59:30 ok... thanks goodness we're out of time 16:59:37 DuncanT-: commented. 16:59:38 please check this wiki later and update 16:59:43 thanks everyone 16:59:47 #endmeeting