17:00:17 <devananda> #startmeeting ironic
17:00:18 <openstack> Meeting started Mon Apr  6 17:00:17 2015 UTC and is due to finish in 60 minutes.  The chair is devananda. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:22 <openstack> The meeting name has been set to 'ironic'
17:00:30 <devananda> #chair NobodyCam
17:00:31 <openstack> Current chairs: NobodyCam devananda
17:00:36 <devananda> hi folks!
17:00:40 <NobodyCam> mornign
17:00:40 <BadCub> heya
17:00:42 <jroll> mornin'
17:00:43 <NobodyCam> :-p
17:00:53 <JoshNang> o/
17:01:08 <NobodyCam> I expect several folks are still on holiday
17:01:13 <aweeks> o/
17:01:21 <rameshg87> o/
17:01:29 <TheJulia> o/
17:01:31 <clif_h> o/
17:01:43 <devananda> as usual, our agenda is posted on the wiki - https://wiki.openstack.org/wiki/Meetings/Ironic
17:01:53 <jlvillal> o/
17:01:56 <devananda> mostly today though, we'll be talking about the release status
17:02:14 <devananda> couple quick announcements / reminders
17:02:17 <JayF> o/
17:02:20 <devananda> #topic announcements / reminders
17:02:28 <devananda> - feature freeze is still in effect
17:02:51 <devananda> - my goal is to tag a release candidate this week, as soon as reasonably possible
17:03:17 <devananda> I'll also be taking a quick holiday thursday and friday, so we kinda need to do it before then :)
17:04:15 <devananda> also, summit planning for all projects has started
17:04:17 <devananda> #link https://wiki.openstack.org/wiki/Design_Summit/Planning
17:04:36 <devananda> (posting that just to make sure everyone's aware)
17:04:51 <devananda> that's it for me -- anyone else?
17:05:04 <BadCub> Drivers page (http://stackalytics.com/report/driverlog?project_id=openstack%2Fironic) needs to be updated so we have an accurate listing
17:05:20 <NobodyCam> will open L spec folder after theats cut?
17:05:30 <NobodyCam> s/will/we'll/
17:05:34 <devananda> oh wait, not quite done - I tagged a 0.5.0 release of python-ironicclient earlier today. I still need to send an announce to the ML (will do after this meeting)
17:05:46 <devananda> NobodyCam: yes, we can open L for both specs and code once the RC is tagged
17:05:56 <NobodyCam> :)
17:06:02 <jroll> woot
17:06:09 <BadCub> devananda: I have opening L tagged on my calendar
17:06:17 <devananda> #info python-ironicclient 0.5.0 tagged today
17:06:35 <NobodyCam> devananda: is there a change log for it?
17:06:45 <devananda> NobodyCam: yah. look at the pip
17:07:00 <NobodyCam> ack :)
17:07:04 <devananda> hm. or the tag here ...
17:07:24 <devananda> #link http://git.openstack.org/cgit/openstack/python-ironicclient/tag/?id=0.5.0
17:07:39 <NobodyCam> ahh TY :)
17:08:00 <devananda> #topic subteam status reports
17:08:37 <devananda> I see a couple notes from IPA and iLO teams that their docs are up and/or need reviews
17:08:45 <NobodyCam> oh cleaning is now enabled?
17:08:52 <devananda> and I see a note from dmitry that discoverd has devstack support
17:08:55 <devananda> NobodyCam: yah
17:09:14 <NobodyCam> :) w00 h00
17:09:37 <NobodyCam> devananda: that was from lastweek but I left it there for this meeting too
17:09:42 <devananda> NobodyCam: ah, ok
17:10:28 <devananda> ok, moving on in a minute ...
17:10:39 <jlvillal> devananda: By docs do you mean specs?
17:10:46 <jlvillal> For Liberty?
17:10:48 <NobodyCam> I've also seen many DOC patches this last week... Awesome !! keep'um comming
17:10:50 <devananda> jlvillal: no, i mean documentation
17:10:51 <jroll> jlvillal: docs.
17:10:55 <jlvillal> devananda: Thanks
17:11:11 <jroll> jlvillal: subteam status reports are at https://etherpad.openstack.org/p/IronicWhiteBoard
17:11:17 <jroll> if you want to follow along
17:11:22 <jlvillal> jroll: Bookmarking
17:11:33 <jroll> (among a ton of other useful stuff)
17:11:53 <devananda> #topic Release Candidate Status
17:12:01 <devananda> #link https://launchpad.net/ironic/+milestone/kilo-rc1
17:12:19 <devananda> that launchpad page is the canonical reference (no pun intended) to the status
17:12:37 <devananda> as this week progresses, either we fix things or we discuss accepting that a bug just won't be fixed in Kilo
17:12:39 <jroll> heh
17:12:40 <devananda> and I untarget it
17:13:03 <devananda> BadCub & others have been keeping track of things in flight on a whiteboard here
17:13:07 <devananda> #link https://etherpad.openstack.org/p/IronicReviewDay
17:14:03 <devananda> we still have a lot of bugs merely in confirmed / triaged status (iow, no one actively working on them)
17:14:03 <NobodyCam> if we have a review that we like eyes on should it be brought up here?
17:14:32 <devananda> I'm inclined to simply defer those bugs now, unless someone wants to take them on today
17:14:38 <devananda> https://bugs.launchpad.net/bugs/1433727
17:14:39 <openstack> Launchpad bug 1433727 in Ironic "partial upgrade not possible, 'reason': u'Unknown argument: "configdrive" (HTTP 400)'" [High,Confirmed] - Assigned to Adam Gandelman (gandelman-a)
17:14:49 <devananda> https://bugs.launchpad.net/bugs/1301279
17:14:50 <openstack> Launchpad bug 1301279 in OpenStack Compute (nova) "Changing node's properties in Ironic after node is deployed will count as available resources in Nova" [Medium,Confirmed] - Assigned to Nha Pham (phqnha)
17:15:01 <devananda> https://bugs.launchpad.net/bugs/1412561
17:15:02 <openstack> Launchpad bug 1412561 in Ironic "IPA/DIB is not compatible to create the uefi capable images" [Medium,Confirmed]
17:15:07 <devananda> https://bugs.launchpad.net/bugs/1177655
17:15:08 <openstack> Launchpad bug 1177655 in Ironic "kernel boot command line for baremetal assumes block device UUID is correct root" [Medium,Triaged]
17:15:12 <devananda> https://bugs.launchpad.net/bugs/1322179
17:15:14 <openstack> Launchpad bug 1322179 in Ironic "Add check on total disk size before deploying" [Medium,Triaged] - Assigned to Sandhya Ganapathy (sandhya-ganapathy)
17:15:17 <devananda> https://bugs.launchpad.net/bugs/1327260
17:15:19 <openstack> Launchpad bug 1327260 in Ironic "Disk layout is not being verified on re-provisioning" [Medium,Triaged]
17:15:23 <devananda> https://bugs.launchpad.net/bugs/1407820
17:15:24 <openstack> Launchpad bug 1407820 in Ironic "server will not boot from pxe if it does not support command 'chassis bootdev pxe ' with 'persistent' option" [Medium,Triaged]
17:15:35 <devananda> https://bugs.launchpad.net/bugs/1326269
17:15:37 <openstack> Launchpad bug 1326269 in Ironic "Developer docs should include information about fake drivers" [Low,Triaged] - Assigned to Shraddha Pandhe (shraddha-pandhe)
17:15:59 <NobodyCam> I'm ok with giving 1326269 more time
17:16:06 <devananda> yea, those two "Low" bugs I'm just bumping
17:16:13 <jroll> the first one seems a bit important (15533727)
17:16:29 <jroll> though if we punt on it, it goes away
17:16:59 <jroll> wow, that number is not at all right
17:17:02 <jroll> 1433727
17:17:11 <NobodyCam> lol
17:17:20 <rameshg87> i would like to bring https://bugs.launchpad.net/ironic/+bug/1422723 to notice
17:17:21 <openstack> Launchpad bug 1422723 in Ironic "Ironic sets instance as active even if bootloader installation fails when boot_option is 'local'" [Critical,In progress] - Assigned to Ramakrishnan G (rameshg87)
17:17:40 <rameshg87> it has tight dependency with it's corresponding dib patch
17:17:43 <NobodyCam> rameshg87: have the DIB changes landed?
17:17:46 <rameshg87> which has already landed
17:17:56 <adam_g> jroll, i think the only solution for that is to document that nova's force_config_drive needs to be False before/during upgrades
17:18:05 <devananda> ouch, yea
17:18:15 <jroll> adam_g: ok, mind putting up a patch for that?
17:18:22 <adam_g> jroll, ya, was planning on doing that
17:18:24 <jroll> rameshg87: so ironic + dib + localboot is broken today?
17:18:27 <jroll> adam_g: awesome, ty
17:18:28 <rameshg87> so in-effect either "ironic + dib should both go in", OR "both shouldn't go in (means reverting back dib patch)"
17:18:32 <rameshg87> jroll: yes :(
17:18:36 <jroll> :|
17:18:41 <devananda> rameshg87: @!(*& :(
17:19:00 <devananda> rameshg87: wait. that's what we use in the gate ... it shouldn't be breakable
17:19:10 <jroll> devananda: we don't use localboot
17:19:12 <rameshg87> devananda: we don't test localboot on gate
17:19:14 <devananda> ahh, right
17:19:28 <jroll> which is silly imo but it is what it is
17:19:46 <devananda> who wants to create a tempest scenario test for localboot?
17:20:00 <devananda> seems like, you know, something we should have done when we added that feature
17:20:06 * rameshg87 will give a try if no one else raises hands before
17:20:11 * jroll nominates lucas :P
17:20:19 <devananda> jroll: lol
17:20:48 <NobodyCam> rameshg87: I'll review 169419 after the meeting
17:20:48 <devananda> so, tempest not testing that aside... i'm concerned with the patch to ironic
17:20:58 <devananda> let's make reviewing that a priority post-meeting
17:21:04 <rameshg87> devananda: yes, please ...
17:21:09 <jroll> +1
17:21:19 <rameshg87> thanks NobodyCam
17:21:26 <NobodyCam> #link https://review.openstack.org/#/c/169419/
17:21:27 <adam_g> does it look like https://review.openstack.org/#/c/164369/ will be making it in this week?
17:21:45 <NobodyCam> thats the one I wanted eyes on
17:22:11 <adam_g> i'd like to get https://review.openstack.org/#/c/167038/ proposed as a stable freeze exception for 2014.2.3 (out thursday), but cant really do that until 164369 is in
17:22:12 <devananda> adam_g: that one's been somewhat contentious
17:22:12 <NobodyCam> the last comment will invoulve hiding more fields?
17:22:52 <devananda> adam_g: I think you can do https://review.openstack.org/#/c/167038/ now
17:23:05 <devananda> adam_g: I just pushed 0.5.0 client release today, which adds the microversion headers by default
17:24:01 <adam_g> devananda, so v1.1 is / will be the equiv of the juno API? it sounded like there was some contention over whether that'd be v1.0 or 1.1
17:24:13 <devananda> adam_g: indeed, there still is contention over that
17:24:29 <devananda> adam_g: or rather, whether NOVERSION = Juno, v1.1 is some point in the beginning of kilo
17:24:39 <devananda> or whether we actually have a supported v1.0 that was Juno
17:25:11 * NobodyCam just want to deside and move on :/
17:25:24 <devananda> I know what I want -- because I proposed it in https://review.openstack.org/#/c/164369/
17:25:51 <jroll> I mean, I don't see anyone that isn't ok with that patch
17:26:00 <jroll> rloo just has some more work to do, seems valid
17:26:04 <devananda> jroll: I haven't seen anyone +2 it
17:26:17 <jroll> sure
17:26:22 <jroll> I need to review it, still
17:26:23 <devananda> which tells me no one is OK with it
17:26:38 <devananda> (no one that reviewed it, anyway)
17:27:01 * NobodyCam is okay with it.
17:27:09 <NobodyCam> :-p
17:27:13 <NobodyCam> *ducks*
17:27:35 <devananda> back-tracking a moment...
17:28:01 <devananda> #info https://review.openstack.org/#/c/169419/ needs priority reviews: critical bug: pxe drivers + dib + local-boot option is broken right now
17:28:53 <devananda> jroll: is anyone going to be NOT-OK if I abandon https://review.openstack.org/#/c/164369/ and we just keep the APIs as they are now?
17:30:12 <JayF> That means custom or non-upgraded clients are broken by default unless they send the correct header, in the general case, yes?
17:30:32 <jroll> devananda: we'll be fine, if that's what you're asking
17:31:27 <NobodyCam> devananda: humm and just leave maintenance reason field exposed?
17:31:27 <devananda> JayF: I dont think so. not sending the header means they get v1.1 API
17:31:33 <jroll> JayF: no, it means things like maintenance_reason etc (that were added before microversions) will be there
17:31:38 <devananda> JayF: which is not exactly the same as Juno, but didn't have any backwards-incompatible changes
17:31:46 <JayF> aha, okay, then I'm okay with that
17:31:47 <jroll> JayF: basically juno vs $(when microversions were added)
17:31:56 <devananda> eg, the NOSTATE->AVAILABLE rename is the really big thing that breaks folks, and that comes after v1.1
17:32:20 <jroll> devananda: theoretically I'd prefer to have that patch, in reality I don't care that much
17:32:26 <jroll> is my opinion
17:32:28 <devananda> so if my patch doesn't land, I think it generally doesn't break anyone. it's just not as _clean_ as I'd like
17:32:48 <devananda> jroll: cool. then it's not worth us discussing much longer, since there's still no concensus on it
17:32:55 * devananda abandons
17:33:06 <jroll> whee.
17:33:27 * NobodyCam closes the ever open tab for 164369
17:34:23 <devananda> here's some more bugs
17:34:31 <devananda> https://bugs.launchpad.net/ironic/+bug/1431929
17:34:32 <openstack> Launchpad bug 1431929 in Ironic "Ironic failing to gracefully handle: ipmi error "insufficient resources for session"" [Medium,In progress] - Assigned to Julia Kreger (juliaashleykreger)
17:34:49 <devananda> TheJulia: you seem to hav ea fix, but you're stuck on unit tests, right?
17:35:10 <NobodyCam> the patch on that is wipped on a timing loop?
17:35:26 <devananda> yea, challenges mocking time calls
17:35:31 <TheJulia> devananda: basically yes.  I've got 3 fo 4 tests working just fine, the ask in a review was to actually mock the timeout
17:35:32 <jroll> quick question on that one: the bug says ipmi tuning didn't help... do we rate limit ipmi commands by node or by host that it connects to
17:35:58 <jroll> IOW, is the real bug that we don't do rate limiting correctly for double bridging?
17:36:00 <TheJulia> jroll: just made it worse :(
17:36:19 <devananda> TheJulia: I'd be fine landing it sans that 4th test then
17:36:29 <TheJulia> I would agree, but really the bug beyond that is that the ipmitool client should be retrying on it's own
17:36:56 <devananda> TheJulia: ipmitool CLI does retry. but it's pretty brutal in my experience
17:37:20 <devananda> i added all the retry / timeout guards in our ipmitool.py module to try to work around issues with ipmitool's CLI
17:37:35 <devananda> https://bugs.launchpad.net/bugs/1436791
17:37:36 <openstack> Launchpad bug 1436791 in Ironic "boot_mode is overwritten in node properties/capabilities when secure boot is requested for iLO drivers" [Medium,In progress] - Assigned to Shivanand Tendulker (shivanand-tendulker)
17:37:36 <TheJulia> yeah, its treating this as a hard failure, hence the retry and string search
17:38:06 <devananda> TheJulia: string search is not really good. what if someone's LOCALE is different?
17:38:17 <jroll> LAST_CMD_TIME.get(driver_info['address'], 0))  # curious if this address is different per node or per bridge, that could be the issue here
17:38:30 <devananda> jroll: it's the same for all devices in a chassis
17:38:32 <jroll> the bridge is getting overloaded because we're spamming it with a thing for each node
17:38:35 <jroll> oh? hrm
17:38:43 <devananda> jroll: right. this code SHOULD guard against it
17:38:50 <jroll> so wth
17:38:51 <TheJulia> devananda: That is a really good point
17:38:57 <devananda> welp. some things are flakey ...
17:39:15 <TheJulia> context seems to be per node, not per bridge address
17:39:21 <devananda> fwiw, I have a new OOB mgmt library in the works which might fix some of these issues for our moonshot
17:39:29 <TheJulia> woot
17:39:31 <devananda> but that's totally off topic. IPMI sanity is important too
17:40:07 <devananda> re the ilo bug above, I suppose we'll leave it to the ilo driver folks, but we should get reviews on that one
17:40:31 <devananda> https://review.openstack.org/167952
17:41:21 <jroll> seems like a lot of code :/
17:42:02 * devananda hasn't reviewed it yet
17:42:19 <devananda> https://bugs.launchpad.net/bugs/1436824
17:42:21 <openstack> Launchpad bug 1436824 in Ironic "Lost control of nodes with amt driver after user reboot " [Medium,In progress] - Assigned to Tan Lin (tan-lin-good)
17:42:38 <devananda> lintan_: how's ^ going?
17:42:57 <rameshg87> devananda: i think code is submitted.  just the doc is pending
17:43:03 <jlvillal> devananda: He might be asleep at this time.
17:43:17 <rameshg87> devananda: https://review.openstack.org/#/c/167975/
17:43:20 <rameshg87> it's merged
17:43:24 <devananda> oh, i see
17:43:27 * devananda closes bug
17:43:35 * rameshg87 notes devananda +Aed :)
17:43:40 <jroll> wait what
17:44:04 <jroll> oh, I see
17:44:05 <devananda> there's a 5-line doc patch pending
17:44:16 <jroll> kind of an odd fix, "only use localboot"
17:44:18 <jroll> but alas
17:45:14 <devananda> https://bugs.launchpad.net/bugs/1271317
17:45:16 <openstack> Launchpad bug 1271317 in Ironic "XML support not required for ironic" [Low,In progress] - Assigned to Devananda van der Veen (devananda)
17:45:27 <devananda> mrda did most of the work, then I redid it this morning
17:45:32 <devananda> should be an easy +2 for folks
17:45:47 <devananda> https://bugs.launchpad.net/bugs/1282836
17:45:48 <openstack> Launchpad bug 1282836 in Ironic "PXE driver does not check if tftpd is running" [Low,In progress] - Assigned to Chris Krelle (nobodycam)
17:46:09 <jroll> this one seems like an optimization to me
17:46:27 <jroll> I guess what I mean is, low-ish value, high amount of distraction
17:46:30 <devananda> NobodyCam: I reviewed ^ this morning. needs a bit of work. It's close to ready, but also very low-impact if we bump it
17:46:31 <JayF> We should be careful that however we fix that honors dhcp_provider=None
17:46:45 <jroll> JayF: +1
17:47:02 <devananda> given that, I'm pretty tempted to bump it
17:47:05 <jroll> TheJulia: you run dhcp_provider=None, but still run tftpd on the conductor yes?
17:47:09 <NobodyCam> devananda: yes and teh dependent patches have string changes
17:47:11 <devananda> jroll: yes
17:47:12 <TheJulia> jroll: yes
17:47:18 <jroll> that's fun.
17:47:42 <NobodyCam> I'm ok with bumping the whole chain
17:47:45 <devananda> k
17:47:45 <jroll> TheJulia: devananda: that makes for some interesting things, we should talk about that more in L
17:48:48 <devananda> NobodyCam: is https://bugs.launchpad.net/bugs/1431929 something we will actually mark as completed soon?
17:48:49 <openstack> Launchpad bug 1431929 in Ironic "Ironic failing to gracefully handle: ipmi error "insufficient resources for session"" [Medium,In progress] - Assigned to Julia Kreger (juliaashleykreger)
17:48:55 <devananda> erm, bad link
17:49:00 <devananda> https://bugs.launchpad.net/bugs/1323589
17:49:00 <openstack> Launchpad bug 1323589 in Ironic "Installation guide needs updating" [Medium,In progress] - Assigned to Chris Krelle (nobodycam)
17:49:04 <devananda> that's the one!
17:49:41 <devananda> NobodyCam: like is there a particular thing wrong in the installation guide that you're fixing?
17:50:02 <jroll> just by the title we can probably leave that open forever
17:50:04 <NobodyCam> I need to see how that got assigned to me
17:50:14 <devananda> oh. huh
17:50:15 <NobodyCam> I did a couple of patches
17:50:17 * rameshg87 sees 10 mins left
17:50:18 <devananda> jroll: indeed
17:50:25 <devananda> NobodyCam: looks like it's old. /me closes
17:50:35 <NobodyCam> ack
17:50:57 <devananda> enough bugs for now -- I'll keep focusing on these for the next couple days
17:51:04 <devananda> and we'll tag RC when the page is clear
17:51:05 <devananda> thanks all!
17:51:09 <devananda> #topic open discussion
17:52:40 <BadCub> https://review.openstack.org/#/c/166192/ - grub2 bootloader support for UEFI Boot Mode was on the list
17:53:15 <rameshg87> BadCub: i guess stendulker is not here
17:53:41 <BadCub> rameshg87: indeed
17:53:44 <rameshg87> BadCub: unfortunately i don't have full details on it :(
17:54:15 <rameshg87> BadCub: but i think he wanted to share some finding on switching default boot loader from elilo to grub2
17:54:32 <rameshg87> BadCub: don't have much info on it
17:54:40 <BadCub> I believe that is correct as well. But neither do I
17:54:43 <NobodyCam> rameshg87: ya that was the concern
17:54:59 <JayF> That seems like a change that shouldn't go into K though, right?
17:55:10 <JayF> Changing bootloaders a few days before RC seems unwise
17:55:23 <rameshg87> JayF: no. i don't think so.
17:55:28 <devananda> JayF: +100
17:55:41 <JayF> Then why is it worth talking about here? Shouldn't it be spec'd for L like anything else going into L?
17:55:44 <NobodyCam> JayF: I think stendulker came up with a way to support both
17:55:57 <BadCub> I believe this was briefly discussed in earlier meetings and decided to hold this off for K.
17:56:57 <devananda> as far as i can tell, this is incomplete feature work from the uefi-secure-boot blueprint
17:57:01 <BadCub> The patch is attached to the BP that was given a FFE for K.
17:57:06 <devananda> meaning that was not actually completed
17:57:43 <BadCub> https://blueprints.launchpad.net/ironic/+spec/uefi-secure-boot
17:57:58 <wanyen> This was a PoC tries to address pxe-i,o secure boot Grub2 migraton concern
17:58:13 <devananda> the FFE was to land the code that was in-flight at that time
17:58:19 <devananda> which was marked completed when https://review.openstack.org/#/c/165907/ landed on 3/19
17:58:27 <BadCub> devananda: correct
17:58:36 <devananda> https://review.openstack.org/#/c/166192/ was proposed a day later
17:58:46 <NobodyCam> * 2 minutes
17:58:57 <devananda> wanyen: if it was known that this was going to be a problem, why wasn't https://review.openstack.org/#/c/166192/ proposed sooner?
17:59:29 <devananda> wanyen: seems like very poor timing to ask for a FFE, then propose new code after that Feature is marked completed ...
17:59:41 <NobodyCam> wanyen: 166192 is only PoC?
17:59:55 <wanyen> devananda, this PoC was implemented after upstream raised concerns about bootloader migration to grub2
18:00:22 <NobodyCam> *time*
18:00:26 <devananda> let's continue this in channel
18:00:35 <NobodyCam> great meeting THank you all
18:00:42 <devananda> thanks all! keep up the great work closing bugs!
18:00:48 <devananda> #endmeeting