17:00:33 <asselin__> #startmeeting third-party
17:00:34 <openstack> Meeting started Tue Dec 22 17:00:33 2015 UTC and is due to finish in 60 minutes.  The chair is asselin__. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:38 <openstack> The meeting name has been set to 'third_party'
17:01:05 <asselin__> anyone here for 3rd party ci working group meeting?
17:01:13 <mmedvede> asselin__: hi
17:01:19 <asselin__> hi mmedvede
17:03:13 <asselin__> ok, I guess just the 2 of us for now...
17:03:17 <asselin__> @link agenda https://wiki.openstack.org/wiki/Meetings/ThirdParty#12.2F22.2F15_1700_UTC
17:03:24 <asselin__> #link https://wiki.openstack.org/wiki/Meetings/ThirdParty#12.2F22.2F15_1700_UTC
17:03:47 <asselin__> #topic announcements
17:04:10 <asselin__> I don't have any announcements this time
17:04:19 <asselin__> mmedvede, you?
17:04:21 <mmedvede> me neither
17:04:38 <asselin__> #topic action items
17:04:54 <asselin__> #link previous action items: http://eavesdrop.openstack.org/meetings/third_party/2015/third_party.2015-12-08-17.05.html
17:05:11 <asselin__> "ci-watch next steps is to add some unit tests"
17:05:18 <asselin__> any updates?
17:05:28 <mmedvede> no
17:05:38 <asselin__> "mmedvede to compare oslotest and testtools"
17:06:12 <mmedvede> so oslotest is mostly a wrapper around testtools, it adds a few features we might find useful
17:06:39 <asselin__> like what?
17:06:47 <mmedvede> it is lightweight. Downside is that it adds extra dependency that we do not need - osclient
17:08:00 <mmedvede> features like ability to timeout a long-running test, tempfile creation, logging configuration
17:08:34 <asselin__> those are nice
17:08:57 <mmedvede> I suggest that because oslotest is already configured to be used in ciwatch through cookiecutter template, we use it
17:09:22 <mmedvede> if we find it useless, can throw it away and just continue to use testtools
17:09:36 <asselin__> how difficult is it to switch between the two?
17:10:32 <mmedvede> does not look hard
17:10:46 <mmedvede> oslotest is still testtools
17:11:06 <ja3> sorry, turns out if I'm not looking at the screen for a while I don't see the reminder pop-ups
17:11:17 <asselin__> ok, I have no objections
17:11:20 <asselin__> hi ja3
17:11:24 <mmedvede> hi ja3
17:12:41 <asselin__> for my action " asselin to created initial blueprints for desired features."
17:13:06 <asselin__> I took a look at launchpad and it requires a bit of pre-setup.
17:13:38 <asselin__> unfortunately I did it after the last meeting and didn't write it down, so I don't remember the details.
17:14:29 <asselin__> so I'll work on that for the next meeting
17:14:45 <asselin__> #topic CI Watch
17:14:57 <asselin__> anything to discuss here?
17:15:43 <mmedvede> asselin__: on this patch - https://review.openstack.org/#/c/258125/
17:17:00 <mmedvede> I was hesitant to +2, as it does more globals. But again, not much can be done before making things like that into classes
17:18:12 <asselin__> yeah I was looking to see how to make them constants, but that's a bit more involved in python.
17:18:41 <mmedvede> It is not really more involved if the codebase was cleaner
17:19:38 <asselin__> do you think we should refactor it into a class with this change or afterwards?
17:20:15 <mmedvede> asselin__: so what do you think, should we do an exception and merge it? The patch definitely makes the code a bit more dirty, even though it is fixing the problem.
17:21:01 <ja3> which is the lesser evil IYO, the degree of incremental dirtniness introduced or getting the problem fixed sooner?
17:21:42 <asselin__> I don't think it's really any more dirty. When we do the refactor, those globals just become part of the class.
17:22:04 <ja3> If smaller easier-to-review patches are preferred, then separating them seems the logical choice.  If you buy separation, then the only question is which order (which nets out to how long are you willing to wait for a Cleaner solution).
17:22:18 <mmedvede> ja3: there is balance. My reason tells me incremental dirtiness is more evil
17:22:48 <ja3> is anyone actually blocked waiting for a fix?
17:23:02 <mmedvede> ja3: no
17:23:09 <mmedvede> not a blocking change
17:23:52 <mmedvede> asselin__: if we agree not to introduce more things like these, I agree
17:24:02 <asselin__> apoorva's not around. we can discuss it offline when he's around
17:24:12 <mmedvede> but if they stay, mere existence of globals would lead to people adding more of them
17:25:27 <asselin__> yes, and globals tend to be inevitable until you have classes
17:27:18 <asselin__> so this is the only outstanding patch. lgtm and focus on unit tests and classes, but of which play hand-in-hand.
17:28:20 <mmedvede> asselin__: there is another one, looks like good improvement
17:28:32 <mmedvede> #link https://review.openstack.org/260193
17:28:41 <asselin__> yeah js change, looking at that now
17:29:22 <mmedvede> pretty straight-forward
17:29:32 <mmedvede> that is it for ciwatch
17:29:54 <asselin__> #topic Common-CI Solution
17:30:18 <asselin__> I've been working on getting the common-ci docs published
17:30:37 <ja3> asselin
17:30:42 <asselin__> #link publish common-ci docs : https://review.openstack.org/#/c/259245/
17:31:09 <asselin__> it's a bit of a big change, but much of it is autogenerated ^^
17:31:25 <ja3> #@$(&# client.  asselin__, I think we'll be ready to consume those in Jan.  I finally am getting clean runs on full-tempest on the z/VM internal CI.
17:31:26 <asselin__> if someone knows a better way to do this, let me know
17:32:47 * mmedvede looking at 259245
17:34:34 <mmedvede> asselin__: it looks like how other projects generate docs, so that should be the way
17:35:23 <asselin__> mmedvede, ok cool thanks
17:35:40 <asselin__> I ran a  recheck which will run the job to generate the docs
17:36:10 <mmedvede> ah, ok. That was my question, because actually publishing the docs is done through jenkins job
17:36:15 <asselin__> once that looks good I'm planning to remove the contrib/README.md file.
17:36:16 <mmedvede> I assume it was added
17:36:53 <asselin__> mmedvede, yes
17:37:09 <asselin__> #link patch to publish common-ci docs once it's merged: https://review.openstack.org/#/c/259587/
17:37:59 <asselin__> #link
17:38:00 <asselin__> Remove logstash/elasticsearch classes, use puppet-openstackci https://review.openstack.org/#/c/240011/
17:38:13 <asselin__> #link Remove logstash/elasticsearch classes, use puppet-openstackci https://review.openstack.org/#/c/240011/
17:38:18 <mmedvede> ja3: that is good to hear
17:38:50 <mmedvede> asselin__: that one I need to look at
17:38:55 <ja3> misha, are you running loganalyze today?
17:38:56 <mmedvede> was in my queue for a while now
17:39:11 <asselin__> mmedvede, whould be nice to get this merged ^^ I think we'll have to ping more infra roots
17:39:13 <mmedvede> ja3: no, we do not run loganalise internally
17:40:17 <mmedvede> asselin__: pls hold on on pinging them, I need to make sure the patch is still valid, and address your comments
17:40:34 <asselin__> mmedvede, yes, of course :)
17:41:11 <asselin__> ok that's it from common-ci
17:41:13 <mmedvede> asselin__: in retrospect, I went the wrong way about that move. It was too big of a change to be made outside of sprint
17:41:48 <asselin__> mmedvede, what does that mean?
17:42:22 <mmedvede> it is hard to get merge, because the original is changing
17:43:15 <mmedvede> there is no freeze. The better strategy is to make changes in system-config first, and then move
17:44:34 <asselin__> but then you have the issue the other way around
17:46:03 <mmedvede> I do not think so. We can discuss it later, I might not be explaining myself clearly :)
17:46:41 <asselin__> sure
17:46:46 <asselin__> #topic open discussion
17:47:01 <asselin__> anything else to discuss from anyone?
17:47:19 <ja3> Just the heads up I pretty much gave before on z/VM's CI status.
17:47:55 <asselin__> ja3, which CI solution is used for that?
17:48:16 <mmedvede> ja3: are you planning to start reporting?
17:48:53 <ja3> We're running the full zuul/etc superstructure, 660 passing 236 skipped by testr 15 skipped by regex (no tempest.conf mechanism we can find... aka patching opp)
17:49:27 <ja3> asselin__, nova and neutron plugins for z/VM which is a mainframe hypervisor... the oldest one actually
17:50:41 <ja3> misha, reporting is going to be a while I suspect.  we're squatting on 2 x86 boxes for everything above z/vm (zuul et al. plus the controller), both the same age (past EOL), and one of them went casters up this month.
17:51:38 <ja3> I had x86 servers already in the pipe, but now we're constrained again.  We could sandbox vote but don't have the width for prod.
17:52:36 <ja3> We also have to procure a log server, which shouldn't be a huge deal I'm starting to work on that now but with the time of year... ugh it probably won't progress too far until Jan.
17:52:56 <asselin__> ja3, where will you host the log server?
17:53:12 <ja3> asselin__, softlayer as a swift server.
17:53:22 <mmedvede> We are using swift container to host logs (currently on SoftLayer)
17:53:46 <mmedvede> ja3: oh, so you are doing the same
17:53:57 <asselin__> my previous team is looking for another place now that hpcloud is going away.
17:54:04 <ja3> we're pretty much required to use softlayer for hosting; and it has several advantages for us actually, like having Someone Else worry about security patching
17:54:19 <mmedvede> +1 for someone else :)
17:54:36 <asselin__> haven't tried swift yet for log server.
17:54:59 <mmedvede> asselin__: it is mostly worry-free. I did have to hack zuul a bit
17:55:10 <ja3> I got instructions from Kurt (about a year old), but of course the process has changed with the intergalactic re-org right after his email.
17:55:40 <asselin__> mmedvede, zuul doesn't support it directly?
17:56:13 <mmedvede> asselin__: it does support it. But it does not support what I wanted to do (change container path a bit). The fix is simple, but still did not merge
17:56:24 <mmedvede> #link improve swift flexibility in zuul https://review.openstack.org/#/c/229582/
17:57:23 <asselin__> thanks, I'll take another look
17:57:32 <asselin__> anything else?
17:57:57 <ja3> not from me
17:58:03 <mmedvede> same here
17:58:14 <asselin__> wish everyone a happy holiday and new year!
17:58:23 <mmedvede> +100
17:58:56 <asselin__> see you next year
17:58:58 <asselin__> #endmeeting