17:07:13 <krtaylor> #startmeeting ironic_qa
17:07:14 <openstack> Meeting started Wed Jul 27 17:07:13 2016 UTC and is due to finish in 60 minutes.  The chair is krtaylor. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:07:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:07:17 <openstack> The meeting name has been set to 'ironic_qa'
17:07:21 <mjturek1> o/
17:07:22 <devananda> o/
17:07:25 <sambetts> o/
17:07:31 <rpioso> o/
17:07:40 <jroll> \o
17:07:50 <krtaylor> hello all, I am assuming that jlvillal is out today
17:08:27 <krtaylor> as always agenda
17:08:32 <krtaylor> #link https://wiki.openstack.org/wiki/Meetings/Ironic-QA
17:08:51 <krtaylor> #topic Announcements
17:08:58 <krtaylor> anyone?
17:09:19 <jroll> sam's big jjb refactor landed \o/
17:09:50 <krtaylor> nice
17:10:03 <jroll> also, infra pointed out to us that we have far more jobs than other projects, and we should be working to reduce that
17:10:11 <sambetts> \o/ and it didn't break everything
17:10:21 <jroll> part of that is our ipmitool transition, which was held back because of reliability concerns
17:10:29 <jroll> I'd like a volunteer to work on the reliability there
17:10:39 <jroll> so we can finish switching over
17:10:46 <clarkb> jjb said 83 integration tests
17:11:15 <sambetts> clarkb: generated or actually being used though?
17:11:25 <clarkb> generated
17:11:30 <clarkb> which is also a problem
17:11:33 <krtaylor> wow
17:12:48 <rajinir> \o
17:13:00 <jroll> right, so we need to work on reducing that
17:13:03 <devananda> I don't have time to dive into the ipmitool proxy work, but I agree it's important to pivot to that soon
17:13:09 <jroll> getting rid of the ssh jobs is one part
17:13:23 <krtaylor> link to share work, spec, blueprint, defect?
17:13:34 <sambetts> https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/projects.yaml#L3212
17:13:50 <jroll> krtaylor: not sure I have one right now
17:13:51 <sambetts> this ^ is were all the different permutations are defined now
17:14:01 <krtaylor> #link https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/projects.yaml#L3212
17:14:07 <krtaylor> thanks sambetts
17:14:08 <devananda> jroll: how about combining the driver tests into a single (slightly longer) run ?
17:14:38 <sambetts> as you can see we have 4 differnt drivers being tested, 2 different types of disk image, then you have src or not-src for the ramdisks
17:14:49 <sambetts> the permutations grow quite fast
17:15:13 <jroll> devananda: yeah, I've thought about doing e.g. whole disk vs partition in one run
17:15:28 <jroll> if we boot two instances on two nodes, and they're configured to use different images
17:15:49 <jroll> seems reasonable for one run
17:17:21 <jroll> so no takers to hack on ipmitool job reliability?
17:17:36 <sambetts> when we remove *_ssh job that'll reduce the generated jobs massivly
17:17:39 <krtaylor> could we get down to testing one (or two) default driver(s)?
17:18:06 <jroll> krtaylor: yes, the goal is to remove the ssh drivers
17:18:15 * sambetts is willing to help with virtualbmc etc
17:18:21 <sambetts> no idea what there is to do though
17:18:21 <jroll> thanks sambetts
17:18:30 <jroll> basically, it's too unreliable to make voting
17:18:33 <jroll> I'm not sure why :)
17:20:10 <sambetts> I've seen weird failures with it even this morning with ironic failing to power on a node, lucas was looking into it
17:21:11 <mjturek1> is there a common failure on the ipmitool job? ie - the one seen here http://logs.openstack.org/77/332177/6/check/gate-tempest-dsvm-ironic-agent_ipmitool-nv/bb13d69/console.html
17:22:32 <mjturek1> just asking because we see this one on the powerkvm ci every now and then and aren't sure why yet ^
17:23:11 <jroll> mjturek1: well, that's a generic "the deploy timed out" sort of thing, the root cause is typically "ramdisk booted too slow or not at all"
17:23:31 <mjturek1> ahhh okay
17:23:38 <rajinir> mjturek1>: Any errors in the conductor logs?
17:24:24 <mjturek1> rajinir: http://logs.openstack.org/77/332177/6/check/gate-tempest-dsvm-ironic-agent_ipmitool-nv/bb13d69/logs/screen-ir-cond.txt.gz?level=ERROR
17:24:31 <mjturek1> yeah it's the node not powering on
17:24:38 <jroll> maybe come back to this in open discussion :)
17:24:39 <mjturek1> which jroll mentioned
17:24:41 <mjturek1> yeah :)
17:25:45 <krtaylor> ok, let's move on then
17:25:56 <krtaylor> #topic Multi-node Grenade
17:26:49 <sambetts> I believe there are a few patches to get this working
17:26:54 <krtaylor> status has not been updated in a while on the etherpad
17:26:56 <jroll> orly?
17:26:59 <sambetts> https://review.openstack.org/335981
17:27:04 <sambetts> https://review.openstack.org/332490
17:27:14 <krtaylor> #link https://etherpad.openstack.org/p/ironic-newton-grenade-whiteboard
17:27:16 <sambetts> https://review.openstack.org/#/c/334382/
17:28:39 <krtaylor> hm, 332490 isn't there - adding it now
17:28:56 <jroll> oh, cool stuff
17:30:40 <krtaylor> anything else to mention here?
17:31:15 <sambetts> No Vasyl has mainly been working on it
17:31:22 <sambetts> and I can't see him online
17:31:55 <krtaylor> ok, lets move on, we can bring it up again in open discussion if needed
17:32:09 <krtaylor> #topic 3rd party CI
17:32:47 <krtaylor> I had an action to create a patch for adding all drivers from a previous meeting, it is here
17:32:50 <thiagop> #info OneView CI is being rebuild due to hosting cloud problems
17:32:50 <krtaylor> #link https://review.openstack.org/#/c/330270/
17:33:16 <krtaylor> thiagop, good to know, thanks
17:33:29 <thiagop> it should be ready already, but I'm having some nasty problems with new versions of jjb, zuul, nodepool...
17:34:13 <krtaylor> thiagop, ping any of the folks in #openstack-third-party-ci for help
17:34:52 <thiagop> krtaylor: thanks, didn't know of this channel..
17:35:47 <krtaylor> jroll, sambetts thanks for the reviews on the drivers patch, I just need the info to fill in the question marks
17:36:14 <krtaylor> wrt the driver vs platform thing, it is really up to the project
17:36:21 <sambetts> Not sure what the best way to handle the drivers on different architechtures is
17:36:30 <jroll> well, imo this should be a list of drivers
17:36:30 <krtaylor> nova lists architectures as different drivers
17:36:40 <jroll> if those drivers are tested on other architectures... we should note that
17:36:42 <jroll> whaaaat
17:37:11 <krtaylor> #link http://docs.openstack.org/developer/nova/support-matrix.html
17:37:12 <jroll> I only see that for kvm_ppc64
17:37:24 <jroll> oh, in their feature matrix, sure
17:37:51 <jroll> but on the driver marketplace, I tend to think that should be a list of possible drivers to be used
17:38:01 <sambetts> ^ this
17:38:20 <krtaylor> and that a driver is supported on a different archtecture?
17:38:26 <krtaylor> thats the question :)
17:38:27 <jroll> put it in the description?
17:38:39 <jroll> "tested on x86_64 and ppc64"
17:38:45 * jroll wrote this in his latest comment
17:38:47 <sambetts> Its a shame we can't list multiple CIs per driver
17:39:06 <sambetts> as far as I can tell anyway
17:39:06 <krtaylor> for community supported drivers tested on other architectures, hm, that works for me
17:39:13 <krtaylor> sambetts, not that I know of
17:39:59 <sambetts> feature request for stacklytics?
17:40:14 <krtaylor> jroll, I think thats a good solution for stackalytics anyway
17:40:29 <krtaylor> we need the contact info for openstack wiki, etc
17:41:08 <jroll> yeah
17:41:44 <krtaylor> I'll push a new rev with that and see what you think
17:42:12 <jroll> awesome, thanks
17:42:33 <krtaylor> anything else for 3rd party CI?
17:42:55 <krtaylor> #topic Open Discussion
17:43:58 <krtaylor> anyone?
17:44:23 <krtaylor> well, looks like we're done then, thanks everyone!
17:44:36 <sambetts> Thanks for hosting krtaylor
17:44:38 <jroll> cool, thanks for hosting krtaylor
17:44:42 <krtaylor> #endmeeting