08:02:41 <anteaya> #startmeeting third-party
08:02:42 <openstack> Meeting started Tue Jun 16 08:02:41 2015 UTC and is due to finish in 60 minutes.  The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:02:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:02:45 <openstack> The meeting name has been set to 'third_party'
08:03:03 <anteaya> if you are here for the third party meeting raise your hand
08:03:06 <anteaya> say hello
08:03:13 <eantyshev> hello!
08:03:56 <anteaya> hello eantyshev
08:04:00 <anteaya> how are you?
08:04:20 <eantyshev> anteaya: Fine, got a couple questions
08:04:29 <anteaya> eantyshev: go ahead
08:06:12 <anteaya> eantyshev: did you want to post your questions?
08:06:22 <eantyshev> current regex for 'recheck' triggers some unrelated CIs, for example https://review.openstack.org/#/c/190273/ patchset 9
08:08:00 <anteaya> yes
08:08:01 <eantyshev> Is it acceptable to narrow the scope of CI to not react to those?
08:08:21 <anteaya> we have no control over what syntax a ci uses for recheck
08:08:35 <anteaya> aslo recheck should fire everything
08:09:01 <anteaya> because otherwise you could just recheck until you got the result you wanted
08:09:10 <anteaya> rather than having all the tests run you just pick
08:09:13 <anteaya> recheck a
08:09:16 <anteaya> recheck a
08:09:20 <anteaya> recheck a
08:09:28 <anteaya> until a passes and then
08:09:31 <anteaya> recheck b
08:09:33 <anteaya> recheck b
08:09:40 <anteaya> until you get the result you wanted
08:09:54 <lennyb> Hi. If recheck fires everything why do we need regex with a proper third party name there ?
08:09:55 <anteaya> this allows patches with race conditions into master
08:10:04 <eantyshev> but 'recheck a' triggers Jenkins jobs too
08:10:15 <anteaya> we never asked for a regex with a third party name
08:10:27 <anteaya> that is just what ci operators did
08:10:38 <anteaya> yes recheck fires jenkins jobs
08:11:03 <lennyb> I see, so basically I can use another keyword to trigger my CI only.
08:11:22 <anteaya> if that is your choice
08:11:37 <anteaya> infra has no enforcement of any recheck syntax
08:11:39 <eantyshev> ok, so it's acceptable for our CI to react to bare 'recheck' and 'recheck Virtuozzo6 CI'?
08:11:49 <anteaya> if that is your choice
08:11:59 <anteaya> infra has no enforcement of any recheck syntax
08:12:25 <eantyshev> what are the reasons to do recheck when it's not requested?
08:12:45 <eantyshev> to show some stability?
08:13:01 <anteaya> recheck means to run the tests
08:13:17 <anteaya> if recheck is the comment running the tests is requested
08:13:27 <anteaya> did you not understand my example above?
08:13:36 <anteaya> you run the tests or you don't
08:13:48 <anteaya> you don't get to chose which tests you run
08:13:54 <anteaya> you run them all
08:14:37 <eantyshev> I think I understand
08:14:41 <anteaya> eantyshev: great
08:15:07 <anteaya> did you have other questions today?
08:15:15 <eantyshev> yes
08:15:56 <eantyshev> many third parties have different accounts for different projects
08:16:03 <anteaya> okay
08:16:41 <eantyshev> does that mean they run separate Zuul server for each?
08:16:46 <anteaya> no idea
08:17:03 <lennyb> we are using 2 zuuls for 2 different CIs
08:17:09 <anteaya> I would guess some do and some don't
08:17:24 <anteaya> we have no way to track who does or does not
08:17:36 <anteaya> lennyb: thanks
08:17:40 <eantyshev> lennyb: do you run them in one VM instance?
08:17:49 <lennyb> I dont think there is a technical problem to use 1 zuuls.
08:17:56 <lennyb> no we use 2 VMs
08:18:08 <lennyb> no. We use 2 VMs
08:18:24 <anteaya> I don't think running more than one zuul on a single server is a good idea
08:18:30 <anteaya> but I don't know for sure
08:18:46 <anteaya> but it wouldn't be an initial choice I made
08:19:11 <anteaya> just like nodepool needs its own network
08:19:20 <anteaya> I would allocate zuul its own server
08:19:36 <eantyshev> lennyb: do you have common jenkins, nodepool, etc. for these CIs?
08:20:13 <lennyb> agreed,  you also use zuul user with it's environment and configuration, so I think it will be very tricky ( if possible ) to configure 2 zuuls on the same machine
08:20:31 <lennyb> we have Jenkins, Zuul, JJB. no nodepool.
08:21:42 <eantyshev> anteaya: we came to the point where we need to split one CI account in 2
08:21:51 <lennyb> btw, we do run NOvaCI and Neutron CI on the same Jenkins Master and zuul, but when our account was disabled due to Neutron CI, Nova CI was disabled as well.
08:21:58 <anteaya> eantyshev: okay
08:23:45 <eantyshev> anteaya: just interested how others do
08:23:57 <anteaya> eantyshev: of course
08:24:08 <anteaya> eantyshev: any other questions?
08:25:15 <anteaya> lennyb: how are you doing today?
08:25:17 <eantyshev> anteaya: have to draw somebody's attention to https://review.openstack.org/188383
08:25:27 <anteaya> lennyb: was there anything you wanted to discuss?
08:25:49 <anteaya> eantyshev: what do you want them to see on 188383?
08:26:17 <lennyb> anteaya: thanks. How are you? BTW, I met jogo yesterday. NO, I have nothing to discuss.
08:26:40 <anteaya> lennyb: wonderful
08:26:54 <anteaya> eantyshev: you see there is a -1 on your patch
08:27:10 <anteaya> eantyshev: jim would like you to provide a test
08:27:18 <eantyshev> anteaya: I want it merged, since all prerequisites are met
08:27:26 <anteaya> eantyshev: you have a -1
08:27:32 <anteaya> and a request for a test
08:27:44 <anteaya> not all prerequistes are met
08:27:54 <anteaya> jim wrote zuul
08:27:59 <eantyshev> anteaya: https://review.openstack.org/191588
08:28:21 <eantyshev> anteaya: it's in dependent review
08:28:35 <anteaya> did you say so in the patch?
08:28:54 <anteaya> have you talked to jim about having the test as a dependent patch?
08:29:16 <anteaya> so you are linking in the meeting so other operators are aware?
08:29:21 <anteaya> that's fine
08:29:36 <anteaya> #link https://review.openstack.org/#/c/188383/
08:29:50 <anteaya> #link https://review.openstack.org/#/c/191588/
08:29:56 <anteaya> thanks for doing thts work
08:29:57 <eantyshev> anteaya: hope this will help, thanks!
08:30:31 <anteaya> eantyshev: I do suggest you find jeblair in infra during north american daylight hours to discuss your approach with him
08:30:47 <anteaya> as he may want the test in the same patch as the code
08:32:15 <eantyshev> anteaya: I will discuss it with him
08:32:22 <anteaya> eantyshev: great
08:32:34 <anteaya> anything else for this meeting?
08:34:36 <lennyb> no.
08:34:41 <eantyshev> no
08:34:42 <anteaya> does anyone else have any items for this meeting?
08:34:45 <anteaya> okay great
08:35:03 <anteaya> well I won't keep you if we are finshed our discussion
08:35:08 <anteaya> thanks for the chat
08:35:11 <anteaya> :)
08:35:18 <lennyb> thanks. buy.
08:35:22 <eantyshev> thanks!
08:35:23 <anteaya> I appreciate your attendance and participation
08:35:24 <lennyb> bye :)
08:35:28 <anteaya> see you next week
08:35:34 <anteaya> #endmeeting