19:01:28 <devananda> #startmeeting ironic
19:01:29 <openstack> Meeting started Mon Sep 22 19:01:28 2014 UTC and is due to finish in 60 minutes.  The chair is devananda. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:29 <JoshNang> o/
19:01:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:32 <openstack> The meeting name has been set to 'ironic'
19:01:37 <devananda> hi!
19:01:45 <NobodyCam> hi all
19:01:57 <devananda> as usual, our agenda is here: https://wiki.openstack.org/wiki/Meetings/Ironic
19:01:59 <jroll> I always feel like this is a talk show with a big musical intro led by the bot
19:01:59 <harshada_kakad> o/
19:02:08 <ChuckC_> hi
19:02:13 <devananda> jroll: lol
19:02:14 <GheRivero> o/
19:02:19 <wanyen_> hi
19:02:23 <rloo> howdy
19:02:33 <devananda> #topic announcements
19:02:53 <devananda> I have a few things that we should probably cover in the opening section
19:02:56 <linggao> Hi
19:03:06 <devananda> apologies in advance for not stashing those on the wiki, they should be quick
19:03:34 <devananda> first, of course, congrats to everyone on all the work getting the project to the point where the TC felt comfortable integrating it
19:03:50 <NobodyCam> great work everyone!!!!
19:03:55 <devananda> that doesn't change any of our plans, and doesn't affect us in the Juno timeframe
19:04:21 <devananda> it does mean we'll actually need a doc writer to be involved in Ironic (and in the OpenStack Docs program) during Kilo
19:04:23 <jroll> \o/ nice work, everyone
19:04:51 <NobodyCam> devananda: will we have to come up with this person? oe will they?
19:04:54 <devananda> that's a significant gap in our current community contributions
19:05:00 <devananda> NobodyCam: who's "they" ?
19:05:13 <NobodyCam> the OpenStack Docs program folks
19:05:14 <devananda> this is not something the openstack doc team provides
19:05:15 <devananda> no
19:05:29 <devananda> we need a person on our team who works with them
19:05:54 <devananda> just like we've had folks on our team who have worked with the qa and devstack and infra folks for the last year
19:05:59 <NobodyCam> do we have a time line to select this person?
19:06:24 <devananda> it could be mroe than one person
19:06:37 <devananda> and I don't think we "select" them
19:06:45 <NobodyCam> :-p
19:06:49 <devananda> anyawy
19:07:18 <devananda> we are still in feature freeze / string freeze / focus on bugs mdoe for a bit longer. and a big thanks to everyone who's working on that
19:07:51 <devananda> also, I am going to run for PTL again this cycle (sending that letter is on my todo for tonight)
19:08:20 * NobodyCam +1's devananda nomanation
19:08:53 <wanyen_> ++
19:08:54 <devananda> I will continue counting on folks to assist with running the project -- and want to thank lucas (who's not here?) and NobodyCam for that
19:09:31 <devananda> in an ideal world, I'd like to have others who could step in for me if need be by the end of Kilo
19:09:31 <jroll> +millions for thanking all three of you
19:09:38 <GheRivero> ++ for all three
19:09:51 <devananda> (I dont like being a SPoF, and I dont think enough folks know how to do the release management things right now, for example)
19:10:15 <devananda> ok - that's all for my announcements!
19:10:20 <devananda> NobodyCam: have anything?
19:10:41 <NobodyCam> nope you covered what I had
19:10:43 <devananda> cool
19:10:48 <devananda> #topic RC status
19:10:49 <jroll> I'd like to announce that CI for the agent driver has landed :D
19:11:02 * jroll will go more into that later
19:11:12 <devananda> jroll: awesome! yep, let's get to that in the IPA section
19:11:15 <devananda> #link https://launchpad.net/ironic/+milestone/juno-rc1
19:11:27 <devananda> we still have a lot of in progress bugs up there
19:11:38 <devananda> 4 are tagged "High"
19:11:54 <devananda> #1279331 conductor lacks periodic task to keep PXE env up to date
19:12:00 <devananda> #1342919 instances rescheduled after building network info do not update the MAC
19:12:06 <devananda> #1347604 ironic-dbsync fails on SQLite due to bad support of migrations
19:12:12 <devananda> #1355510 conductor hash ring is not updated as nodes arrive and leave
19:12:51 <Shrews> devananda: 1347604 is just documentation now, so not really "high" at this point
19:13:05 <devananda> #1279331 and #1355510 are related. last week we talked a bit about that -- what's the sense on whether we can make those changes in Juno?
19:13:05 <dtantsur> I'd still call it high
19:13:22 <wanyen_> deva, https://bugs.launchpad.net/ironic/+bug/1370466 is not on the RC1 milestone.  Can ywe add it on the list?
19:13:26 <uvirtbot> Launchpad bug 1370466 in ironic "pxe_ilo driver will fail on ipmitool error after switching from uefi to bios boot mode" [Medium,In progress]
19:13:48 <devananda> Shrews, dtantsur: I don't think a doc bug is high at this point in the release cycle -- it's not something to block the release on.
19:14:35 <dtantsur> devananda, well, it make our quick start guide not work as it's written. don't know if it's high
19:14:38 <devananda> wanyen_: it looks like a patch is in flight, though i need to review the bug's alternatives
19:14:54 <wanyen_> deva, yes the fixes has been submitted.
19:15:05 <devananda> dtantsur: not saying it's not a bug -- just that that doesn't meet the standards for "high"
19:15:21 <dtantsur> probably, yeah
19:15:27 <Shrews> dtantsur: we could close it with my current patch, and open a new bug for documenting ironic-dbsync more thoroughly
19:15:33 <dtantsur> right
19:15:35 <devananda> wanyen_: targeted
19:15:46 <wanyen_> deav, ty
19:15:56 <wanyen_> s/deva/deva
19:16:02 <devananda> so, what's the concensus on the hash ring?
19:16:07 <devananda> i think we need to make a decision today
19:16:23 <devananda> and either fix it in the next few days or bump it to kilo
19:16:24 <jroll> "use the agent driver" is my answer to that :P
19:16:31 * NobodyCam would vote to bump to kilo
19:16:33 <jroll> my question is, do we think we can fix it that quickly?
19:16:46 <Shrews> devananda: greghaynes suggested bumping, and i agree
19:16:50 <devananda> if we bump it to kilo, we need to include documentation about that limitation in the release notes
19:16:51 <dtantsur> I don't think we can fix it in a couple of days
19:16:52 <rloo> doesn't even look like anyone is assigned to 1279331
19:16:56 <JayF> +1 to release note
19:17:27 <rloo> would love to have hash ring fixed in next few days, but not sure it is possible. is there a (sorry) design for it?
19:17:34 <NobodyCam> I think that is the best course at this late stage in the release cycle
19:18:07 <NobodyCam> bumpping to kilo that is
19:18:22 <devananda> rloo: afaik, no formal spec. lots of bugs and discussion on them, though
19:18:26 <jroll> +1, it's sad but I think that's the best route
19:18:41 <devananda> yea, that makes me a bit sad too
19:18:54 <rloo> devananda: so the bugs/discussions (for me) ought to be agreed on first, before reviewing any code.
19:19:16 <devananda> i should have kept better track of that work post midcycle
19:20:02 <devananda> it's a fairly serious bug for anyone using the curent pxe driver and more than one conductor, fwiw
19:20:32 <jroll> agree
19:20:38 <jroll> and I do expect to see people deploying juno
19:20:40 <rloo> devananda: there is more than RC1 before the juno release. do we have more than a few days to fix the hash ring?
19:20:49 <devananda> are there other bugs which folks feel should block the release?
19:21:18 <dtantsur> about "#1342919 instances rescheduled after building network info do not update the MAC"
19:21:33 <dtantsur> I think it should be clarified in the bug, what else needs to be done
19:21:37 <jroll> dtantsur: I thought that was fixed?
19:21:53 <dtantsur> it looks like blocking rc1, but does not look like it requires changes in Ironic itself
19:21:57 <devananda> rloo: yes. RC1 is the point we re-open the main development tree
19:22:09 <devananda> rloo: but we can continue to have RC2...N if necessary over the next few weeks
19:22:11 <jroll> I'm fairly certain that's done, hm
19:22:23 <devananda> rloo: the problem is, they take significantly more work to backport bug fixes to those branches
19:22:26 <dtantsur> jroll, that's why I'm asking for clarification
19:22:39 * jroll looks
19:22:40 <rloo> devananda: so I think we should try to get the hash ring stuff in.
19:22:40 <devananda> rloo: so if we plan on making some big change / fixing soem big bug, it's better to delay the RC1 tag
19:22:48 <devananda> rloo: I agree
19:22:58 <jroll> https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L538-544
19:22:59 <rloo> oh, if we can delay RC1, that might be better. How long can we delay it for?
19:23:02 <devananda> I would actually really like to see it fixed in Juno
19:23:03 <jroll> this was the fix afaik
19:23:08 <jroll> dtantsur: ^
19:23:37 <jroll> comstud: https://bugs.launchpad.net/ironic/+bug/1342919 is fixed here, yes? https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L538-544
19:23:39 <uvirtbot> Launchpad bug 1342919 in ironic "instances rescheduled after building network info do not update the MAC" [High,In progress]
19:24:29 <devananda> rloo: we /should/ tag it this thursday. delaying it means delaying downstream testing
19:25:34 <Shrews> devananda: hrm, if the hashring stuff is going to affect usability, we probably *should* try to get something in
19:26:01 <rloo> devananda: so maybe we should discuss after meeting, about hash ring. i don't know what the status is with that. I'm willing to help out, but I think we need to see where we're at and what needs to be done to fix.
19:26:16 <devananda> ok. let's dig into that after the meeting
19:26:33 <devananda> if there are no other bugs that folks feel should block RC1, I will start bumping things tomorrow
19:26:47 <devananda> and the release will be tagged when https://launchpad.net/ironic/+milestone/juno-rc1 contains no open bugs
19:26:57 <dtantsur> I'm really feeling like closing https://bugs.launchpad.net/ironic/+bug/1342919 right now. Objections?
19:26:58 <devananda> I will bump anything below High at this point, fwiw
19:26:59 <uvirtbot> Launchpad bug 1342919 in ironic "instances rescheduled after building network info do not update the MAC" [High,In progress]
19:27:26 <devananda> dtantsur: oh - taht's fixed in Nova now?
19:27:46 <devananda> if that's a bug in the virt driver,a nd fixed in Nova, it's not a bug in Ironic any more, right?
19:28:06 <dtantsur> devananda, jroll posted https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L538-544
19:28:18 <dtantsur> which presumably means it's fixed
19:28:21 <NobodyCam> would it be worth trying to get bugs like: https://bugs.launchpad.net/bugs/1323589 thru
19:28:22 <uvirtbot> Launchpad bug 1323589 in ironic "Installation guide needs updating" [Medium,In progress]
19:28:32 <devananda> NobodyCam: no. that can be done later
19:28:37 <NobodyCam> ack
19:28:39 <devananda> NobodyCam: doc bugs don't block a code release
19:28:50 <NobodyCam> ++
19:28:55 <devananda> since committing changes to the doc/source/* tree immediately updates the website
19:29:02 <devananda> there's no need to backport those to a release candidate branch
19:29:17 <devananda> dtantsur: closing that bug looks appropirate
19:29:24 <dtantsur> ack thanks
19:29:29 <devananda> ok. going to move on so that we have time to cover other things too
19:29:29 <dtantsur> -1 high bug!
19:29:43 <devananda> #topic Kilo summit planning
19:29:50 <devananda> keeping this -really- brief today
19:30:13 <wanyen_> deva, we have not submitted ilo driver doc and uefi doc.  Do we need to file a bug for them?
19:30:24 <devananda> just a reminder that, once the RC is tagged, we'll open development and start accepting specs and thinking abotu planning and stuff
19:30:33 <devananda> wanyen_: those don't need bugs. please submit the docs ASAP.
19:30:42 <wanyen_> deva, ok
19:30:46 <NobodyCam> do we have any ideas for the pod signup sheet? or just in general how the will be handled
19:31:10 <NobodyCam> ...how the POD will be...
19:31:19 <devananda> NobodyCam: I don't know what day(s) we'll be scheduled on, or hwo many official sessions we'll have.
19:31:32 <devananda> NobodyCam: so yes, we'll need to run a POD, and we'll need a process for that.
19:31:40 <devananda> I've suggested using a google doc
19:31:43 * devananda finds the link
19:32:27 <devananda> #link https://docs.google.com/spreadsheets/d/1XBKdeDeGfaRYaThjIIoYRwe_zPensECnxsKUuqdoVmQ/edit#gid=0
19:32:37 <devananda> please start jotting down your session ideas there
19:32:40 <devananda> it's open to anyone to edit
19:32:47 <devananda> we'll discuss it in the following weeks (but not today)
19:33:27 <devananda> also, fyi, if you want to discuss something in another program's track
19:33:27 <NobodyCam> ack
19:33:31 <devananda> those are linked / tracked here:
19:33:32 <devananda> #link https://wiki.openstack.org/wiki/Summit/Planning
19:33:44 <devananda> some are using etherpads, some are using gdocs...
19:33:59 <devananda> there will be no planning website like we had at the last developer summits
19:34:04 <devananda> moving on ...
19:34:11 <devananda> #topic subteam reports
19:34:28 <adam_g> heya!
19:34:34 <devananda> adam_g: hi! testing status?
19:34:51 <adam_g> hi, still having a hard time getting -infra to push two changes that will enable the grenade testing.
19:34:51 <devananda> adam_g: looks like we're still waiting for the tempest parallel -nv test to be merged to infra
19:35:00 <devananda> hrm
19:35:24 <adam_g> jroll has landed all of the IPA stuff and it looks like there is a new non-voting job that will be propagating out to jenkins slaves now
19:35:40 <NobodyCam> w00t
19:35:44 <devananda> I think we're going to need the grenade test to be run on the RC branches to give us confidence that it will pass on the juno/stable branch once that is cut
19:35:56 <devananda> jroll: awesome!
19:36:19 <devananda> adam_g: anything I can do to help?
19:36:30 <adam_g> devananda, yeah. ill be more vocal today about getting that stuff reviewed
19:36:40 <adam_g> seems i can get one +2, then it languishes for a rebase before the next core will look at it
19:36:46 <devananda> :-/
19:36:48 <jroll> adam_g: btw, can you add the agent_ssh job to http://no-carrier.net/~adam/openstack/ironic_gate_status.html ? :)
19:37:00 <adam_g> jroll, yea
19:37:00 <jroll> (no rush)
19:37:03 <jroll> thanks
19:37:11 <adam_g> should probably make that page a bit more readable
19:37:31 <adam_g> aside from the on-going CI stuff, one point i wanted to bring up re stable/icehouse
19:38:09 <adam_g> the current tempest job is voting on stable/icehouse patches, but our API has changed in juno and the baremetal tests no longer pass against incehouse
19:38:32 <devananda> adam_g: oh? what part of our API ?
19:38:44 <adam_g> due to addition get/set boot device API, and some changes around console
19:38:58 <devananda> I thought branchless tempest' goal was to prevent that sort of thing
19:39:04 <devananda> (but IMBW)
19:39:06 <adam_g> yeah, it is
19:39:12 <devananda> oooh - i see
19:39:21 <adam_g> but we haven't been paying too close attention to the stable/icehouse branch
19:39:22 <devananda> we added features to the API without making tempest able to NOT run those tests
19:39:29 <adam_g> the features were added, then tests
19:39:49 <devananda> adam_g: yup. thanks for bringing that up. we'll need to keep track of stable/juno much better during Kilo than we did with stable/icehouse
19:40:11 <devananda> adam_g: are there bugs open for this?
19:40:23 <adam_g> my question: should we turn tempest off for stable/icehouse and maintain that branch as a best effort, and promise API stability starting with juno?
19:40:36 <devananda> it seems like we should make the client smart enough to discover that those endpoints aren't present, rather than what it is probably doing now
19:40:50 <mtreinish> devananda: well you should be voting in kilo, so those should changes should be blocked across a release boundary
19:41:10 <adam_g> devananda, well, we're using the tempest baremteal client, not python-ironicclient.
19:41:34 <devananda> adam_g: right. we need to do both, IMO.
19:41:55 <devananda> mtreinish: i'm not sure what you mean by "blocked" here...
19:42:16 <mtreinish> it would bounce of the gate with a failing test if you proposed a breaking api change
19:42:17 <devananda> mtreinish: taht further changes of this type wouldn't be possible because stable/juno will vote on kilo patches in Ironic and tempest, or something else?
19:43:00 <devananda> mtreinish: if we proposed a new test in tempest which passed on master/kilo but failed on stable/juno?
19:43:18 <mtreinish> devananda: yeah basically the branchless tempest advantages should prevent you from breaking an api that's already got coverage
19:43:46 <devananda> mtreinish: great
19:44:05 <adam_g> devananda, as i said, one issue was the new features added. the other was a straight up modification to the API: https://review.openstack.org/#/c/87760/
19:44:05 <mtreinish> devananda: well yeah it'll still be dependent on test coverage, but: http://docs.openstack.org/developer/tempest/overview.html#branchless-tempest-considerations
19:44:51 <devananda> adam_g: ah. yup. that happened :(
19:45:13 <devananda> adam_g: so we should prevent that happening again
19:45:42 <devananda> moving on so we dont run out of time
19:45:43 <devananda> dtantsur: bugs! we covered the release blocking ones already, but do you have anything to add?
19:45:45 <adam_g> devananda, definitely, and i think we will, since stable/juno will be better maintained and see official releases
19:45:53 <dtantsur> numbers!
19:45:59 <dtantsur> Open: 100 (-11). 3 new (-3), 32 in progress (-4), 0 critical (-1), 10 high (-1) and 4 incomplete (-1); juno-rc1: 15 (-3)
19:46:11 <dtantsur> good job, though 15 for rc1 is hard ;)
19:46:33 <dtantsur> also, I started again pocking folks about forgotten bugs
19:46:51 <devananda> dtantsur: thanks!
19:46:52 <dtantsur> please have a look at what you're assigned to:
19:46:53 <NobodyCam> dtantsur: high is at 9 now?
19:47:08 <dtantsur> #link http://ironic-bugs.divius.net/
19:47:09 <devananda> yea, we have 9 High bugs, but only 4 targeted to RC
19:47:21 <NobodyCam> ;)
19:47:52 <devananda> dtantsur: I'm going to bump non-High RC-targeted bugs in the next day or so, fyi
19:47:56 <dtantsur> heh, my site is sometimes slow to update :)
19:47:57 <devananda> drivers -- jroll, linggao, and wanyen -- anything to announce or discuss?
19:48:04 <dtantsur> ack
19:48:08 <jroll> hi!
19:48:18 <jroll> as I said before, we now have upstream CI for agent_ssh driver
19:48:42 <jroll> now the main focus for IPA, for the remainder of the cycle, is docs and packaging
19:48:56 <jroll> JayF: did we make any decisions this morning about packaging IPA?
19:48:57 <devananda> jroll: awesome
19:49:07 <JayF> Would it be OK if we punted on real "packaging" of IPA this cycle?
19:49:14 <devananda> JayF: ++
19:49:15 <linggao> Yes, we have added a "IBM xCAT CI" in Ironic review page for the third party CI test for your patches.
19:49:16 <wanyen_> we have landed iso image build script to ipa.
19:49:21 <JayF> Tag a release in git, yes. Upload to pypi, yes... but maybe instruct distro packager to hold off a cycle
19:49:23 <jroll> JayF: I'm cool with that :)
19:49:32 <wanyen_> we are working on doc and do more testing
19:49:41 <linggao> It tests the ipminative with real hardware.
19:49:46 <jroll> JayF: and by packaging I mean any/all packaging, not only distro. but I like your plan.
19:49:47 <devananda> wanyen_: nice!
19:49:49 <linggao> It is non-voting
19:49:55 <jroll> linggao: that's awesome :D I'd like to do the same
19:49:57 <NobodyCam> wanyen_: awesome:)
19:50:03 <JayF> devananda: wdyt? Agreed with upload to pypi and tagging in git but punting on distro packaging?
19:50:06 <wanyen_> we plan to submit dib element for iso image build script
19:50:15 <devananda> wanyen_: what are your thoughts on getting third-party CI system reporting tests of the iLO driver on real hardware?
19:50:31 <wanyen_> we are lookin ginto that for kilo
19:50:59 <jroll> wanyen_: CI can be set up any time, even during feature freeze :)
19:50:59 <devananda> JayF: I think it's too late tojust now  start packaging for a Juno release
19:51:29 <devananda> linggao: great job on the third-party CI stuff. I've added a little info to the wiki, but I would appreciate it if you have more you can share
19:51:43 <linggao> ^_^
19:51:51 <devananda> #link https://wiki.openstack.org/wiki/Ironic/Drivers
19:51:58 <JayF> devananda: curious what you mean, in detail, by packaging... would my proposal of cutting a release to pypi and tagging something in git be OK?
19:52:01 <jroll> ^ I added to that already for agent things
19:52:41 <devananda> JayF: let's talk a bit more, I think it's worth getting infra's opinion on it as I'm not a packaging/release expert
19:53:04 <JayF> okay
19:53:17 <devananda> JayF: fwiw, I think pushing a tagged release to pypi is the right approach, but IMBW
19:53:58 <devananda> linggao: I would love to have a page that shows me the status of all ibm-xcat tests over time, similar to the page adam_g has here for the upstream tests
19:54:06 <devananda> linggao: http://no-carrier.net/~adam/openstack/ironic_gate_status.html
19:54:31 <devananda> eventually I would like a summary of all the test stats to be on the wiki
19:54:55 <linggao> devananda, ok. We'll see what we can do for that.
19:55:02 <NobodyCam> five minutes
19:55:04 <devananda> thanks!
19:55:15 <wanyen_> Just to clarify- we did ilo driver test on real hardware
19:55:33 <devananda> any other comments or questions about drivers?
19:55:46 <devananda> will move on to open discussion in 20 seconds :)
19:56:07 <devananda> #topic open discussion
19:56:11 <NobodyCam> wanted to ask if we should come up with a version / feature comparison chart. I've already started to questions about what packaged version of openstack work with Ironic...
19:56:42 <jroll> NobodyCam: "packaged version" being like the things HP / mirantis package up and ship?
19:56:51 <NobodyCam> *started to SEE questions
19:57:10 <jroll> NobodyCam: I would think the answer would be "anything that runs on juno" for now
19:57:39 <NobodyCam> and the distros like the zigo stuff this mornign in channel
19:57:41 <jroll> that said, I don't think, as a project, we should care about packaged versions of openstack; those questions should be pointed at $packager support
19:57:44 <jroll> ah ok
19:57:49 <devananda> NobodyCam: so, we can not and should not get in the habit of answering questions about distros
19:57:56 <devananda> jroll: exactly
19:58:18 <NobodyCam> ack ... just want to check
19:58:22 <NobodyCam> wanted
19:58:33 <devananda> NobodyCam: zigo's questions seemed to boil down to "does current trunk of Ironic work with the .deb of Icehouse Nova?"
19:58:43 <devananda> NobodyCam: which is quite different from a version / feature comparison chart
19:58:55 <devananda> which some other openstack projects have tried to maintain
19:59:01 <devananda> like, what features work on vendor X hardware
19:59:09 <devananda> I actually really want us to avoid that
19:59:24 <jroll> 30 seconds
19:59:28 <NobodyCam> ac ++
19:59:43 <NobodyCam> thank you all! great meeting
19:59:47 <jroll> let's move to our channel, thanks everyone :)
19:59:53 <devananda> cheers, thanks all!
19:59:54 <wanyen_> tx
20:00:02 <devananda> #endmeeting