16:00:00 <mhayden> #startmeeting openstack_ansible_meeting
16:00:01 <openstack> Meeting started Thu Mar  2 16:00:00 2017 UTC and is due to finish in 60 minutes.  The chair is mhayden. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:03 <xgerman> o/
16:00:05 <openstack> The meeting name has been set to 'openstack_ansible_meeting'
16:00:08 <mhayden> right at 16:00
16:00:10 <mhayden> such joy
16:00:14 <mhayden> #topic Roll Call
16:00:21 <alextricity25> o/
16:00:23 <xgerman> o/
16:01:15 <andymccr> o/
16:01:20 <andymccr> you are the best at timing mhayden
16:01:36 <xgerman> +1
16:01:59 <jmccrory> o/
16:02:03 <rromans> o/
16:02:21 <palendae> o/
16:02:42 <mgariepy> o/
16:03:03 <d34dh0r53> o/
16:03:15 * mhayden waits 'til 16:05
16:03:17 <spotz> \o/
16:05:12 <mhayden> okay, let's get rolling
16:05:24 <mhayden> last week was the PTG, so there were no action items ;)
16:05:34 <logan-> o/
16:05:35 <mhayden> #topic Operations Guide Review - asettle
16:05:41 * mhayden hands it over to asettle
16:05:43 <asettle> Hey dueds
16:05:47 <asettle> ... a new version of dudes
16:05:51 <mhayden> sup dued
16:05:52 <asettle> Well, first up
16:06:01 <asettle> THanks for those that helped out reviewing the guide at the PTG - solid effort
16:06:14 <asettle> Secondly - there's something I should have made clear during that effort
16:06:37 <asettle> Anything that isn't *specific* OSA config and operations content - remove it. But don't delete - we'll move this to docs.o.o
16:07:01 <asettle> Has anyone been able to see if anyone can help contribute?
16:07:10 <asettle> I know evrardjp has offered to write a section (thanks man!)
16:07:24 <andymccr> i think i signed up for a section too
16:07:40 <mhayden> woot
16:08:01 <andymccr> all the cool kids are signing up for ops guide sections these days... you should all try it
16:08:07 <mhayden> hah
16:08:21 <mhayden> asettle: got a link to the remaining work?
16:08:27 <evrardjp> a new one?
16:08:29 <asettle> uno momento
16:08:49 <asettle> #link https://wiki.openstack.org/wiki/OpenStackAnsible#Documentation_planning
16:08:58 <evrardjp> if all the cool kids are doing it, I'm out.
16:09:13 <asettle> andymccr: could you put it in the table plz? :)
16:09:18 <andymccr> evrardjp: you were one of the cool kids
16:09:20 <asettle> Just so we can track what needs to be done
16:09:26 <evrardjp> andymccr: that's great!
16:09:31 <asettle> I appreciate you're not all operators, but, you have that knowledge :)
16:09:32 <asettle> So, yay
16:09:32 <evrardjp> #hasbeenalready
16:09:33 <andymccr> asettle: im in the table bottom section!
16:09:36 <jmccrory> what's the difference between removing and deleting?
16:09:40 <asettle> andymccr: oh that guy
16:09:51 <asettle> jmccrory: removing (commenting out), deleting (out of tree and out of mind)
16:09:55 <andymccr> hard to figure out based on my irc nick but i am infact "Andy McCrae"
16:10:01 <jmccrory> gotcha
16:10:02 <spotz> hehe
16:10:02 <crushil> \o
16:10:05 <crushil> sorry I'm late
16:10:11 <asettle> Wow andymccr tell me more
16:10:15 <asettle> I just... I feel like I don't know you very well
16:10:24 <mhayden> crushil: welcome!
16:10:25 <asettle> crushil: nobody will every forgive you
16:10:29 <asettle> ... two sides of OSA
16:10:32 <asettle> ever*
16:10:35 <crushil> lol
16:10:39 <andymccr> crushil: welcome!
16:11:00 <asettle> crushil: are *you* an operator? Do you do operations?
16:11:00 <odyssey4me> o/
16:11:01 <spotz> asettle is vicious this week:)
16:11:11 <asettle> spotz: it's not been a fun week :P
16:11:16 <asettle> I"M SORRY ILL BE NICER
16:11:25 <palendae> asettle: Don't succumb to peer pressure
16:11:37 * asettle runs around in circles
16:11:40 <asettle> What do I do, what do I do
16:11:51 <spotz> Pick on Nova more?:)
16:12:09 <asettle> spotz: I am like, the last in a very long line of people hatin' on nova this cycle :P
16:12:11 * asettle points of andymccr
16:12:13 <asettle> at*
16:12:15 <asettle> Frick I just can't type
16:12:19 <asettle> mhayden: we are done with the ops update :P
16:12:24 <mhayden> woot!
16:12:45 <spotz> If you couldn't tell...:)
16:12:49 <mhayden> #topic Python project for inventory code - palendae
16:12:51 * mhayden hands it to palendae
16:13:31 <crushil> asettle, not really. I was an operator in my previous life. Trying to be part of a contributor in multiple projects including OSA
16:13:42 <palendae> There's a review up at https://review.openstack.org/#/c/418076/ to move the inventory code into a Python package. Would appreciate input on naming. stevelle already provided some feedback
16:14:22 <palendae> My main concern with naming it - it may grow to have more than inventory stuff in it, so I'd opt for not referencing only inventory
16:14:29 <palendae> (and yes, I will fix the merge conflict today)
16:14:39 <palendae> #link https://review.openstack.org/#/c/418076/
16:14:43 <palendae> And that's all I have
16:15:04 <asettle> crushil: that's alllll I need ;)
16:15:08 <asettle> https://wiki.openstack.org/wiki/OpenStackAnsible#Documentation_planning
16:15:11 <asettle> sorry palendae
16:15:16 <andymccr> palendae: so the aim is to review that PR once its unmergeconflicted
16:15:35 <palendae> andymccr: That and provide names...maybe I should set up a place to vote separately
16:15:59 <andymccr> palendae: ok cool
16:16:30 <andymccr> palendae: so we'd have to move that out into a separate repo (once it's named) i assume?
16:16:35 <odyssey4me> palendae do we intend to move it out into its own repo to become a library?
16:16:45 <palendae> I don't intend to move it yet, no
16:16:58 <palendae> You can have more than one package in a repo; glance already does this
16:18:07 <andymccr> palendae: i guess once the bits are there we can figure out the details afterwards?
16:18:14 <palendae> Yeah
16:18:24 <andymccr> sounds good to me.
16:18:27 <palendae> They don't strictly need to exist in a separate repo to be a pip-installable package
16:18:43 <palendae> So we can get something installable for now and I guess if people really want more repos we can
16:19:06 <odyssey4me> ok cool, keeping it in-repo certaonly makes testing easier
16:19:27 <andymccr> yeah i think its a future looking thing, we could do at some point if/when it makes sense but for now it doesnt really make sense
16:19:34 <andymccr> (splitting out into its own repo)
16:19:38 <odyssey4me> and if we can do more than one package from a single repo, then there is hope for one day doing 'pip install openstack-ansible' and magic happens!
16:19:46 <andymccr> odyssey4me: magic already happens
16:19:54 <cloudnull> o/
16:19:54 <palendae> odyssey4me: Right. But with a different name :p
16:20:02 <andymccr> No matching distribution found for openstack-ansible
16:20:05 <andymccr> magic!
16:20:58 <andymccr> ok we good on this topic? palendae mhayden?
16:21:00 <evrardjp> pip install osa
16:21:11 <evrardjp> should do everything for us
16:21:19 <evrardjp> that's magic.
16:21:26 <palendae> https://etherpad.openstack.org/p/osa-python-package-name for voting I guess
16:21:28 * mhayden yields to palendae if we're done discussing this one
16:21:54 <palendae> With that link, I think we can be :)
16:22:09 <andymccr> #link https://etherpad.openstack.org/p/osa-python-package-name
16:22:12 <evrardjp> https://martinfowler.com/bliki/TwoHardThings.html
16:22:15 <andymccr> #action vote for a name!
16:22:24 <mhayden> woot
16:22:25 <palendae> evrardjp: indeed
16:22:31 <mhayden> some voting (and trolling) already underway
16:23:07 <andymccr> we can probably collect a list and have a vote at next meeting?
16:23:13 <andymccr> unless actual votes go into the etherpad
16:23:15 <mhayden> sounds good
16:23:23 <palendae> andymccr: Works for me
16:23:23 <mhayden> palendae: get something out to the ML for that if you can
16:23:29 <mhayden> okay, moving right along
16:23:37 <mhayden> #topic Release Planning & Decisions - andymccr
16:23:40 * mhayden passes it to andymccr
16:23:48 <mhayden> i hear there's a big release on the way
16:23:51 <mhayden> something about a deadline
16:23:53 <andymccr> you heard correctly!
16:24:06 <andymccr> so next week is the deadline, i'll be doing the release either tomorrow or early next week.
16:24:12 <andymccr> What i need from everybody is
16:24:17 <andymccr> 1. tell me any critical blockers
16:24:29 <cloudnull> erator for the win  !
16:24:30 <cloudnull> imo
16:24:51 <andymccr> 2. help review https://review.openstack.org/#/c/439691/ - which is a blocker
16:24:59 <spotz> If I'm not about ping for reviews
16:25:05 <andymccr> jamesdenton raised some upgrade bugs which that should fix and i'd like that in for release
16:25:14 <andymccr> but if there are no other blockers i'll do a sha bump once that's in and release straight away
16:25:38 <spotz> andymccr: Did you verify against the comment from Matt? I looked at it yesterday but wasn't sure you were going to change things again
16:26:36 <andymccr> spotz: yeah there is a tricky setup in that new builds vs upgrades have a different process
16:26:51 <spotz> You're all approvedL)
16:26:54 <andymccr> which is why the patch is needed :) but yeah
16:26:55 <andymccr> thanks!
16:27:13 <andymccr> newton/mitaka releases - i need to check if it goes out this week or next
16:27:26 <andymccr> i'll keep with the 2 week cadence, so if we released 2 weeks ago this will be the next release week
16:27:38 <andymccr> again any issues or questions let me know!
16:27:51 <andymccr> and yeah - key thing - notify me of any blockers for the release!
16:28:30 <andymccr> ok all done on release update.
16:28:31 <mhayden> okay, good to move along?
16:28:34 <andymccr> mhayden: si si
16:28:46 <mhayden> i'm going to skip down a little
16:28:55 <mhayden> #topic ceph_client on 16.04
16:29:04 <mhayden> anyone want to speak to this item? there was no name on it
16:29:26 <andymccr> logan-: perhaps?
16:29:36 <logan-> wasn't me
16:29:47 <mgariepy> maybe cloudnull ?
16:29:54 <logan-> i'm not sure, i'm using it and no huge issues afaik
16:29:56 <evrardjp> I think the idea
16:30:08 <evrardjp> was to have deployer's opinions
16:30:08 <cloudnull> not i
16:30:27 <evrardjp> for the ceph_client thing. There is supposed to have two links there for the meeting
16:30:54 <evrardjp> two weeks ago this was added to the agenda due to a conversation during triage
16:30:59 <evrardjp> let me give you the links
16:31:28 <evrardjp> so that's the bug:
16:31:30 <evrardjp> https://bugs.launchpad.net/openstack-ansible/+bug/1661948
16:31:30 <openstack> Launchpad bug 1661948 in openstack-ansible "ceph_client: on ubuntu 16.04 default Ceph version (hammer) causes conflicts " [Undecided,New]
16:31:32 <evrardjp> #link https://bugs.launchpad.net/openstack-ansible/+bug/1661948
16:31:44 <evrardjp> and that's our conversation
16:31:46 <evrardjp> #link http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/%23openstack-ansible.2017-02-07.log.html#t2017-02-07T16:15:24
16:32:02 <evrardjp> we were waiting two weeks ago for a feedback from Travis IIRC
16:32:22 <evrardjp> or for any other deployer
16:33:40 <evrardjp> so do we do this in this "stable" branch ?
16:33:54 <andymccr> hnn
16:34:01 <evrardjp> or
16:34:17 <evrardjp> we just leave it as is, and wait for another input.
16:34:30 <andymccr> i think we need more input tbh
16:34:46 <andymccr> i know logan- and a few others are actively using ceph - and if we can't re-create the problem it'll be hard to fix it also
16:34:53 <evrardjp> logan-: what's the version currently deployed?
16:35:27 <evrardjp> also keep in mind that now we have uca everywhere, and not only "base xenial repos"
16:35:28 <logan-> newton is still hammer by default I think
16:35:52 <evrardjp> I suggest we keep it that way, and wait for more feedback.
16:36:03 <evrardjp> oh I already said taht
16:36:13 <evrardjp> I'm becoming a goldfish
16:36:20 <spotz> UCA stuff has been rock solid, they put up a working liberasurecode1 for me in netwon-proposed if anyone needs it
16:36:23 <logan-> hmm actually ceph_stable_release is not defined in newton, that's strange
16:37:13 <logan-> oh it's in vars
16:37:15 <logan-> yeah hammer
16:37:25 <evrardjp> it's in vars
16:37:47 <logan-> ocata+ is jewel by default, so we should see this problem less over time as we maintain the ceph version better
16:37:54 <evrardjp> ok
16:38:13 <logan-> i will bump ceph_client to the next ceph LTS in Q so we don't get behind like this again
16:38:28 <evrardjp> if we see the bug appearing on N, we'll think of a bump of ceph version and a bump of osa version for N branch I guess :)
16:38:36 <evrardjp> logan-: great!
16:38:52 <logan-> actually pike might be best, i'm not sure, but depending on the timing of the LTS
16:39:01 <evrardjp> If that's a periodic thing, maybe you can add that to our openstack-ansible agenda :)
16:39:16 <logan-> ok
16:39:53 <mhayden> are we good on this topic?
16:40:04 <evrardjp> I think so
16:40:25 <logan-> yeah
16:42:12 <logan-> #action logan- add ceph_client version bump to development cycle checklist
16:42:30 <evrardjp> #link https://docs.openstack.org/developer/openstack-ansible/ocata/developer-docs/contribute.html#development-cycle-checklist
16:42:41 <evrardjp> this way it's logged next to action :)
16:43:19 <mgariepy> the ceph client version should be compatible with older release no ?
16:43:21 <mhayden> okay, movin' along
16:43:35 <mhayden> #topic Bugs needing attention - evrardjp
16:43:40 <evrardjp> mgariepy: that's a good point
16:43:43 <mhayden> evrardjp: anything pressing this week in bug land?
16:43:43 <evrardjp> oh ok
16:43:50 <evrardjp> yes
16:43:56 <evrardjp> We have 2 Critical bugs:
16:44:07 <evrardjp> Very slow Xenial gate jobs and Rebuilding keystone[0] container breaks credential keys
16:44:22 <evrardjp> the first one is sitting for a while, and the second need some love I guess
16:44:28 <andymccr> i'd also like to add this one as a key bug:
16:44:29 <andymccr> https://bugs.launchpad.net/openstack-ansible/+bug/1667103
16:44:29 <openstack> Launchpad bug 1667103 in openstack-ansible "Error upgrading MariaDB during N->O upgrade" [High,Confirmed]
16:44:46 <logan-> for the keystone bug, I plan to have a patch up tomorrow and it will need backports and added to releases asap
16:44:52 <mhayden> slow xenial jobs are a PITA for sure
16:44:54 <evrardjp> I was planning to go to high bugs later, but that's fine
16:45:06 <evrardjp> we have 4 high level bugs
16:45:08 <mhayden> canonical is still working on the xenial perf problems
16:45:23 <evrardjp> I think these bugs need either work, or re-classification
16:45:44 <andymccr> logan-: if its up tomorrow we can perhaps wait for that before releasing?
16:46:16 <evrardjp> I'd be enclined to wait for the critical to merge before releasing
16:46:20 <logan-> yep
16:46:25 <andymccr> although i guess that impacts other releases so i wouldnt say its a release blocker atm
16:47:47 <logan-> yeah i wouldn't say it is a blocker but it is something we should get in asap if the fix is accepted quickly
16:47:54 <andymccr> logan-: agreed
16:48:02 <evrardjp> it's not a blocker by itself but it definitely decrease our deployer experience and we may look bad if things are broken are passed into a tag
16:48:21 <evrardjp> so the faster the fix with a tag, the better, IMO
16:48:25 <andymccr> it seems it impacts mostly on newton when you would do the OS upgrade
16:48:35 <andymccr> but yeah agreed we need that fixed and included
16:48:39 <evrardjp> oh ok
16:48:43 <evrardjp> thanks for clarificaiton
16:49:02 <evrardjp> I think that's all I wanted to say, bugs need more love :)
16:50:22 <mhayden> alrighty
16:50:26 <mhayden> #topic Open Floor
16:50:33 <mhayden> we have ~ 9 minutes
16:50:43 <jmccrory> started notes on newton-xenial upgrade here https://etherpad.openstack.org/p/osa-newton-xenial-upgrade
16:50:51 <jmccrory> still need to add computes and do more testing
16:50:57 <evrardjp> thanks jmccrory!
16:51:00 <jmccrory> but anyone else that's tried and has input, feel free to add to the page
16:51:25 <jmccrory> and move to docs whenever it's looking reasonable
16:52:01 <andymccr> jmccrory: sounds good
16:52:17 <andymccr> logan- was going to be adding some notes anyway perhaps collab on that etherpad would be good
16:52:17 <evrardjp> I think the "fresh install of xenial" is really important, ask mhayden :D
16:53:01 <mhayden> trusty -> xenial upgrades are TURRIBLE for me so far
16:53:13 <jmccrory> yep, ubuntu's script has been finicky
16:53:33 <evrardjp> mhayden: could you add your comments on point 5 on the etherpad?
16:53:39 <evrardjp> :)
16:53:42 <mhayden> can do
16:54:14 <evrardjp> cool thanks
16:54:53 <mhayden> okay, going to close this up unless there's anything else
16:55:20 <andymccr> all good this side
16:58:24 <evrardjp> yup
16:58:27 <evrardjp> thanks everyone!
16:59:04 <mhayden> #endmeeting