15:00:18 <anteaya> #startmeeting third-party
15:00:20 <openstack> Meeting started Mon Oct 12 15:00:18 2015 UTC and is due to finish in 60 minutes.  The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:25 <openstack> The meeting name has been set to 'third_party'
15:00:27 <anteaya> hello
15:00:36 <eantyshev> hello!
15:00:43 <anteaya> hello eantyshev
15:00:48 <anteaya> how are you today?
15:01:23 <eantyshev> rather good, thank you!
15:01:28 <anteaya> glad to hear it
15:01:38 <anteaya> is there anything you would like to discuss today?
15:01:47 * mmedvede lurking around
15:01:54 <anteaya> hello mmedvede
15:02:03 <eantyshev> last week I tried several times to 'promote' my zuul change: https://review.openstack.org/207094
15:02:04 <anteaya> no point in lurking with just the three of us
15:02:07 <mmedvede> hi everyone
15:02:16 <mmedvede> good point
15:02:26 <anteaya> eantyshev: when you say promote what do you mean?
15:02:33 <eantyshev> could be of particular interest for 3rd party CIs
15:03:04 <eantyshev> I mean to draw attention of somebody able to merge
15:03:13 <lennyb> Hi
15:03:19 <asselin_> hi
15:03:23 <anteaya> eantyshev: ah okay great
15:03:28 <anteaya> lennyb asselin_ hello
15:03:44 <anteaya> eantyshev: well I think you had quite a few conversations with clarkb about it last week
15:03:52 <anteaya> so you are getting some attention
15:04:13 <mmedvede> eantyshev: I am interested in the change. We currently have 600 "fake" projects in our zuul layout just because of that problem
15:04:22 <anteaya> we are discussing this patch
15:04:23 <eantyshev> yes, but it seems to get distracted and there was no conclusion
15:04:25 <anteaya> #link https://review.openstack.org/#/c/207094/
15:04:38 <anteaya> eantyshev: yes, that is the world of infra
15:05:11 <anteaya> I suggest that after today try to find jeblair and see if you can get him to leave a comment on the patch with his thoughts
15:05:46 <anteaya> changes to queuing behaviour are made reluctantly since we don't want to break what is working
15:05:47 <eantyshev> mmedvede: it would be great to get your thoughts on that too
15:05:54 <clarkb> eantyshev: I wanted you to talk to jeblair about it
15:06:00 <anteaya> clarkb: thank you
15:06:01 <clarkb> so no conclusion but a definitive next step
15:06:19 <mmedvede> eantyshev: I do not know why I did not leave +1 on it, reviewing now
15:06:43 <anteaya> eantyshev: so clarkb gave you a next step, can you follow through on that?
15:07:00 * asselin_ will also review
15:07:03 <eantyshev> clarkb: sorry, I thought you wanted to discuss this with him by yourself
15:07:32 <eantyshev> clarkb, anteaya: I'll contact jeblair when I see him
15:07:58 <anteaya> eantyshev: wonderful, since parts of the states are having a holiday today, you may find him today you may not
15:08:02 <anteaya> any more on this topic?
15:08:30 <anteaya> does anyone have another topic they would like to discuss today?
15:08:34 <lennyb> yep
15:08:39 <anteaya> lennyb: go ahead
15:08:39 <lennyb> #link https://review.openstack.org/#/c/231439/
15:08:51 <anteaya> what is the context please?
15:09:20 <lennyb> sometimes we need to combine more than 1 run of tempest to results
15:09:34 <anteaya> can you expand with details?
15:09:42 <anteaya> I can't envision what you describe
15:10:09 <lennyb> this patch cambines N subunit files to 1 html with backward capatibility to all CIs
15:10:28 <anteaya> does anyone else understand what lennyb is trying to do
15:10:34 <anteaya> and can try to explain it to me
15:10:39 * lennyb writing
15:11:26 <asselin_> anteaya, so if you have to run multiple different invocations of test to run different unit tests, how can you combine the results of each invocation into a single html report
15:11:34 <lennyb> Sometimes we need to run external script between tempest tests and we need to combine tempest output files ( subunit file ) to one html result
15:11:37 <anteaya> asselin_: thank you
15:13:16 <anteaya> okay I suggest you find folks who are core reviewers on that repo and have a chat with them about your patch
15:13:33 <anteaya> I don't know enough about testr and the vision for it to form an opinion
15:13:43 <anteaya> but thank you for taking the time to offer a patch
15:14:08 <lennyb> anteaya: sure. I wander if any other CI ops need to combine tempest runs to one html result file
15:14:16 <anteaya> good question
15:14:27 <anteaya> anyone else experiencing the same issue as lennyb?
15:15:16 <clarkb> just combine the stream files?
15:15:25 <lennyb> clarb: yes
15:15:27 <mmedvede> lennyb: not us. I see the usefulness of it though
15:15:59 <asselin_> I also see if useful. Can the files not be combined outside of the script though?
15:16:23 <clarkb> lennyb: no I mean just use cat
15:16:30 <clarkb> I am not sure why os-testr needs to support this directly
15:17:05 <lennyb> asselin_: yes, basically they can be.
15:18:19 <lennyb> clarkb: yes, they can, but I've thought it will be nice to expand functionality of the script :)
15:19:20 <clarkb> generally I prefer to leave that ot the shell
15:20:53 <asselin_> I agree with clarkb. If it can be done separately, it follows the shell pipe-and-filter design paradigm
15:22:10 <anteaya> lennyb: can you understand the feedback you have gotten?
15:22:20 <anteaya> do we need more time on this topic?
15:22:23 <mmedvede> wouldn't you have to do some sed/gluing of several html files then?
15:22:27 <lennyb> anteaya: yes
15:22:34 <anteaya> lennyb: great thank you
15:22:40 <anteaya> can we move on?
15:23:35 <anteaya> let's move on
15:23:49 <anteaya> does anyone have another topic they would like to discuss today?
15:24:38 <mmedvede> so I have a simple review as well (as I see we bring reviews up), for those interested in zuul swift upload
15:24:50 <mmedvede> #link https://review.openstack.org/#/c/229582/
15:25:40 <anteaya> mmedvede: what would you like folks in the meeting to do?
15:25:41 <mmedvede> It rearranges code and makes zuul swift upload a bit more configurable
15:25:50 <anteaya> are you asking for help in getting it moving?
15:26:04 <mmedvede> anteaya: not asking for help moving
15:26:09 <anteaya> okay
15:26:20 <anteaya> are you asking for folks to review it?
15:26:28 <anteaya> or just making them aware the patch exists?
15:26:48 <mmedvede> yes, but only those who are interested in zuul swift upload
15:27:01 <anteaya> okay for those interested, this patch exists
15:27:05 <anteaya> any comments?
15:28:10 <anteaya> does anyone have anything more on this topic?
15:28:36 <anteaya> any objection moving to another topic?
15:28:53 <mmedvede> go ahead, thank you
15:28:58 <anteaya> thank you mmedvede
15:29:10 <anteaya> does anyone have any other topic they would like to discuss?
15:29:55 <anteaya> does anyone have any objection to me closing the meeting?
15:30:07 <asselin_> I have something
15:30:14 <anteaya> asselin_: go ahead
15:30:35 <asselin_> my single-node common-ci patch merged last week
15:30:42 <anteaya> congratulations
15:30:46 <anteaya> nice work
15:31:07 <asselin_> #link single-node common-ci https://review.openstack.org/#/c/200330/
15:31:38 <anteaya> well done
15:31:57 <asselin_> Thanks! I'm still working on adding some documentation
15:32:26 <anteaya> has anyone had a chance to test asselin_'s work at all?
15:32:30 <asselin_> although that patch has quite a bit already that should be sufficient for those already familiar with puppet & 3rd party ci
15:32:50 <eantyshev> asselin_: does it allow to create over-cloud on many machines?
15:32:56 <mmedvede> asselin_: nice. I might try to use it next time we need to create a separate pipeline, easier than bringing separate servers
15:33:08 <anteaya> #link single-node common-ci documentation-patch https://review.openstack.org/#/c/227584/
15:33:16 <asselin_> eantyshev, what do you mean by "over-cloud"?
15:33:57 <asselin_> mmedvede, cool, and thanks for your reviews on that patch, they were very helpful!
15:33:58 <eantyshev> asselin_: nova instance from which to get slave VMs
15:34:22 <asselin_> eantyshev, yes, the patch uses nodepool, and you can configure nodepool to connect to many clouds
15:34:30 <mmedvede> asselin_: yw, glad to be helpful
15:35:36 <eantyshev> asselin_: I saw some nodepool plugin for devstack, and thought it allows to establish the cloud
15:36:12 <asselin_> eantyshev, I think that one is used for testing purposes
15:36:42 <asselin_> eantyshev, otherwise, you can configure nodepool to connect to any cloud vi it's auth url, username and password
15:37:04 <asselin_> eantyshev, in our setup, we connect to 5-6 of them
15:37:29 <eantyshev> asselin_: thanks, I got it
15:37:41 <anteaya> nice work here asselin_
15:38:08 <anteaya> so if you know someone starting out perhaps you can suggest they try puppet-openstackci
15:38:15 <asselin_> #link nodepool example: http://git.openstack.org/cgit/openstack-infra/project-config-example/tree/nodepool/nodepool.yaml#n44
15:38:23 <asselin_> eantyshev, see this example ^^
15:38:29 <anteaya> and if they hit a blocker help them figure out how to file a bug
15:39:01 <asselin_> eantyshev, however I will be migrating it to use the oscc, clouds.yaml file, this way there are no credentials in nodepool.yaml
15:39:02 <eantyshev> asselin_: nice thing to start with, indeed
15:39:20 <asselin_> eantyshev, but otherwise this is what we use currently
15:39:43 <asselin_> eantyshev, just add another section to "providers" for each cloud you need
15:40:15 <eantyshev> asselin_: I already use clouds.yaml
15:40:26 <asselin_> eantyshev, ok, then you're one set ahead :)
15:40:30 <asselin_> step
15:40:43 <mmedvede> asselin_: can you remind the gist of why it is a bad thing to have credentials in nodepool.yaml
15:41:13 <eantyshev> asselin_: and should I move on to use secure.conf, BTW?
15:41:15 <mmedvede> because then you can not store it in project config?
15:41:38 <mmedvede> asselin_: I think I answered my own question :)
15:41:40 <asselin_> mmedvede, it's not necessarily 'bad', it just means you need to restrict access to the project that contains your passwords
15:42:08 <asselin_> eantyshev, yes you need to. The single-node ci scripts will automatically set that up for you
15:43:15 <eantyshev> asselin_: I don't feel like to completely switch my CI to the new puppets
15:43:58 <eantyshev> asselin_: so, just making changes to the private puppets tree
15:45:16 <asselin_> eantyshev, up to you of course. However you should at least use the puppet-openstackci classes
15:46:39 <anteaya> is there more to be discussed on this topic?
15:47:41 <anteaya> thank you for mentioning the topic asselin_
15:47:55 <asselin_> if anyone tries to use the patch or has questions, please ping me
15:48:03 <anteaya> did you have plans to post to the -dev mailing list?
15:48:07 <anteaya> asselin_: will do
15:48:19 <anteaya> or did you want the docs patch to merge first?
15:48:29 <asselin_> anteaya, yes, and yes
15:48:39 <anteaya> ah great
15:48:42 <asselin_> but open to opinions
15:48:45 <anteaya> sounds good then
15:48:56 <anteaya> having docs in place first will save you a load of time
15:49:21 <anteaya> very nice work here
15:49:25 <asselin_> thanks
15:49:27 <anteaya> this is a great achievement
15:49:30 <anteaya> :)
15:49:40 <anteaya> *round of applause*
15:49:49 <mmedvede> +1
15:50:13 <anteaya> does anyone have anything else they would like to discuss today?
15:50:16 <asselin_> thanks...glad to finally get to the point :)
15:50:22 <asselin_> this*
15:50:29 <anteaya> yes, it has been a long haul
15:50:36 <anteaya> thank you for the sustained effort
15:51:57 <anteaya> any more topics to discuss today?
15:52:27 <anteaya> any objection to me closing the meeting?
15:52:59 <anteaya> thank you all for your kind attendance and participation today
15:53:41 <anteaya> as advance notice the third party meetings I chair, this one Monday utc 15:00 and Tuesday utc 08:00 will be canceled the week of summit
15:53:58 <mmedvede> anteaya: thanks
15:54:02 <anteaya> tonyb is working on a new function in yaml2ical to skip meetings on the agenda
15:54:06 <asselin_> anteaya, thans
15:54:10 <anteaya> thank you
15:54:15 <anteaya> see you next week
15:54:21 <anteaya> #endmeeting