17:00:43 <tdawson> #startmeeting CentOS PaaS SIG
17:00:43 <centbot> Meeting started Wed Oct 25 17:00:43 2017 UTC.  The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:43 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:00:50 <tdawson> #topic roll call
17:00:51 * _ari_ here
17:00:54 <_ari_> made it back
17:01:00 * jdetiber waves
17:01:07 <jdetiber> will be afk for a couple of min though :)
17:01:15 <tdawson> hi jdetiber
17:01:33 <tdawson> jdetiber: OK, if you're afk when we get to multi-arch, we'll do some other stuff first.
17:01:37 <tdawson> Hi _ari_
17:01:44 * _ari_ waves
17:01:48 <tdawson> _ari_ glad you made it back in time
17:02:38 <_ari_> me too
17:03:32 * jdetiber is back
17:04:33 <tdawson> Does anyone know if sdodson is in today?  I'm not seeing him in chat.
17:04:57 <_ari_> no idea?
17:05:25 <tdawson> #chair tdawson jdetiber _ari_
17:05:25 <centbot> Current chairs: _ari_ jdetiber tdawson
17:05:38 <tdawson> Well, we'll start without him.  People tend to get in late anyway.
17:05:45 <tdawson> #topic OpenShift on CentOS Current Status
17:05:50 <tdawson> Starting off with rpms
17:06:13 <tdawson> I don't have any information ... all I have I told last time.
17:07:01 <tdawson> We have 3.7.0 alpha1 built, and it's in all the correct repo's including origin37
17:07:11 <tdawson> And the corresponding openshift-ansible
17:07:40 <tdawson> Does anyone know if there is an alpha2 that needs to be built?
17:08:13 <tdawson> Moving on to automation
17:08:19 <tdawson> _ari_: How are things on that front?
17:08:39 <_ari_> tdawson: excellent I have https://github.com/CentOS-PaaS-SIG/paas-sig-ci/pull/42
17:09:09 <_ari_> tdawson: ran it through some testing https://ci.centos.org/job/paas-ci-cd/
17:09:19 <_ari_> It runs the whole pipeline
17:09:36 <tdawson> Very cool
17:09:38 <_ari_> Tasks after that would be the following:
17:09:38 <_ari_> - build an openshift cluster and execute e2e tests (done by other team member just needs to be merged)
17:09:38 <_ari_> - move everything to run inside openshift on containers instead of using duffy.
17:09:38 <_ari_> - Figure out how we also run on bare metal for multi-arch, but Jason and I can work that out.
17:09:56 <tdawson> _ari_: So it will do both "future" as well as the release?
17:10:06 <_ari_> As I mentioned in email to you is - he part I am working on now is a better way to anticipate a new tag and then automatically create a release.  I think I have a method just in the middle of testing it out.  This would allow us to continuosly build on master for origin and openshift-ansible and than anticipate a release.
17:10:08 <_ari_> yes
17:10:29 <tdawson> _ari_ That is aweseome
17:10:39 <_ari_> exactly so we would have one job and it would run differently depending on the data it will have a trigger job to call it
17:10:44 <tdawson> _ari_: And ya, I did read it in the email, just wanted to make sure it got said here too. :)
17:11:04 <_ari_> tdawson: yes np just restating it for the broader audience
17:11:19 <_ari_> tdawson: that is all from me for automation
17:12:06 <tdawson> _ari_: That is some tremendous work, thank you for that.
17:12:21 <tdawson> Does anyone have any questions before we move on to multi-arch ?
17:12:34 <_ari_> tdawson: np it has been fun to do I will also take the libs I have created and add them to cico shared library as well
17:13:08 <tdawson> OK, moving on to multi-arch
17:13:21 <tdawson> jdetiber: Are you still back?  If so, how are things in the multi-arch universe?
17:13:51 <jdetiber> So, I was able to get the multi-arch job I'm using for testing to work with a trigger on merges to origin master
17:14:06 <jdetiber> But, it quickly exhausted available disk space on both our slave and then on the jenkins master
17:14:23 <jdetiber> So, it's currently disabled, while I work on refactoring it a bit to make use of the artifact host
17:14:48 <tdawson> jdetiber: Is it a bug in the code, or do we need to add hardware?
17:15:15 <jdetiber> In other news, bstinson has informed me that aarch64 instances in duffy are almost ready to go
17:15:31 <tdawson> Ya!! go bstinson :)
17:15:36 <jdetiber> tdawson: basically, too many copies of the origin code :)
17:16:00 <tdawson> jdetiber: Ya, it is rather large, especially if you have the whole git repo
17:16:19 <jdetiber> there is a repo checkout for each workspace and data is rsync'd back and forth between the slave and the duffy instance, so it all added up
17:16:25 <tdawson> trimming out .git helps a bit, but also hurts, depeneding on what you are doing.
17:16:37 <jdetiber> and I was probably a bit greedy about storing too much in the job artifacts
17:16:44 <tdawson> :)
17:17:34 <tdawson> hi sdodson
17:17:35 <jdetiber> Also, I had a meeting with some folks in the CentOS and Fedora communities about a multi-arch OpenShift cluster for ci purposes. I will be sending out a summary to the mailing list later today
17:17:55 <sdodson> hey, sorry i guess this channel isn't in my auto join
17:18:15 <tdawson> jdetiber: Very cool .. I'm sure alot of people would want a multi-arch cluster.
17:18:34 <tdawson> sdodson: No problem, though you did miss _ari_'s wonderful report on the automation.
17:18:35 <jdetiber> tdawson: not to spoil the surprise, but we are probably talking about multiple clusters :)
17:18:46 <tdawson> Even better. :)
17:18:59 <tdawson> s390x too?
17:19:05 <tdawson> Sorry, just had to ask that.
17:19:13 <jdetiber> tdawson: well, we need s390x support in CentOS for that first ;)
17:19:37 <tdawson> I'll probrubly never touch a s390x in my lifetime, it's just sorta fun to say.
17:19:59 <tdawson> jdetiber: Understood
17:20:02 <jdetiber> heh, at some point I'll need to visit one of our datacenters and get a picture next to one :)
17:21:45 <tdawson> jdetiber: Anyway, I might have derailed what you were saying.  That's pretty exciting about the multi-arch clusters.
17:21:55 <jdetiber> As a tangential aside, I'm shadowing the CI Signal and Upgrade Testing roles for the k8s 1.9 release. As part of that I'm hoping to work on expanding a bit of the alt arch testing and trying to get better alt-arch coverage as part of the release process
17:23:04 <_ari_> jdetiber: should we join forces on that?
17:23:15 <_ari_> Did you get a chance to look at my PR?
17:23:37 <jdetiber> _ari_: I have not yet, ETOOMANYDISTRACTIONS
17:23:49 <tdawson> I was going to say, it looks like you two are sorta working on the same things.
17:23:49 <_ari_> jdetiber: :)
17:23:55 <jdetiber> _ari_: but yes to joining forces :)
17:24:08 <_ari_> jdetiber: I don't want to duplicate effort
17:25:05 <jdetiber> _ari_: so far, I've been focusing mostly on validation, unit, and integration testing, hoping that we can insert that into the pipeline prior to your testing
17:25:32 <jdetiber> _ari_: plus, I'm hoping that having a multi-arch cluster for CI use will help with extending your tests for multi-arch :)
17:25:35 <_ari_> jdetiber: so your pipeline would trigger my pipeline then
17:25:44 <jdetiber> _ari_: basically, yes :)
17:25:45 <_ari_> cool
17:26:47 <jdetiber> tdawson: I think that covers multi-arch, unless bstinson has anything to add
17:27:16 <tdawson> Very nice, thank you both for your reports, and all the work you're doing.
17:27:17 <bstinson> all good here :) thanks for doing that meeting the other day jdetiber
17:27:36 <tdawson> Moving on to Documentation
17:27:48 <jdetiber> bstinson: np, will have the summary ready to send this afternoon
17:28:09 <tdawson> I haven't had to update anything, and it sounds like things are *almost* ready to put links to documentation, but not quite.
17:29:38 <tdawson> So, unless someone has anything to link to, we'll mov on to Images and Image building
17:30:10 <tdawson> sdodson: Since it's been a couple of weeks, just pinging you to know if you have any image news.
17:32:43 <tdawson> Anyone have anything else to input about images and / or image building?
17:33:29 <tdawson> I'm also not finding anyone for minishift or kompose
17:33:46 <tdawson> Well, then let's move to OpenFloor
17:33:53 <tdawson> #topic Open Floor
17:34:11 <tdawson> Does anyone have anything that we haven't already covered?  Or that we might have covered and missed?
17:34:12 <sdodson> tdawson: sorry, no update on those images
17:34:29 <sdodson> tdawson: The one thing I had to mention today is we need to get updated 3.6 installer built
17:34:36 <sdodson> tdawson: is that something I can be shown how to do;/
17:34:38 <sdodson> ?
17:35:06 <tdawson> _ari_: Can we do that with the automation?  Do just openshift-ansible?
17:35:29 <_ari_> tdawson: we built 3.7 didn't we already build 3.6
17:35:45 <tdawson> _ari_: We did, but we need to update the installer for 3.6
17:36:24 <_ari_> tdawson: not sure what you mean if we generate a new openshift-ansible aren't we all set
17:37:04 <_ari_> A more general question is recently skuznets and justin showed their CI/CD is there a reason they don't do this?  Are they only focused on a subset
17:37:14 <tdawson> _ari_: We need a new one off the 3.6 branch, instead of master
17:37:45 <kstange> is it known that rh-php71 is not able to be satisfied due to missing dependencies?
17:38:06 <kstange> it seems to be partially on the public mirrors
17:38:19 <tdawson> sdodson: I still have my manual steps I can get you.
17:38:41 <_ari_> tdawson: I can build it with the automation
17:38:59 <_ari_> I think we need to separate the origin and openshift-ansible a bit
17:39:29 <tdawson> _ari_: Ya, that might be a good idea.
17:39:51 <tdawson> Maybe have some type of checkbox for origin, openshift-ansible, or both.
17:40:04 <_ari_> tdawson: exactly
17:42:24 <_ari_> tdawson: so openshift-ansible 3.6
17:42:36 <_ari_> and official cbs
17:42:42 <tdawson> _ari_: Yep
17:44:32 <_ari_> tdawson: ok will do that no
17:44:32 <_ari_> w
17:44:46 <tdawson> _ari_: Thank you very much
17:45:20 <tdawson> Anything else before we end the meeting for the week?
17:45:42 <_ari_> still no work from brenton sdodson
17:46:41 <kstange> to amend my previous question, I see that the rh-php71 packages do exist on buildlogs, is there anything that needs to be done to help them along to a release?
17:48:06 <tdawson> kstange: I'd send a quick email to centos-devel.  The rhscl people are usually pretty responsive there.
17:48:10 <kstange> I'm sorry, I didn't realize this was a meeting that i'm talking over
17:48:20 <kstange> tdawson: thanks
17:48:24 <tdawson> kstange: No problem, we're towards the end.
17:49:51 <tdawson> kbsingh: Maybe it would be better to try going through you to ask OpenShift to supply some more people resources to the Paas SIG.
17:50:23 <tdawson> I'm sure that sdodson has asked, but maybe hearing it from two fronts would be bettter.
17:52:06 <_ari_> tdawson: just curious since they showcased CI/CD for online but maybe that is separate although these are openshift bits
17:52:10 <_ari_> just curious
17:52:40 <sdodson> tdawson, _ari_ -- justin and brenton were supposed to follow up
17:56:08 <tdawson> sdodson: OK, we'll follow up with them next week.
17:56:27 <tdawson> Looks like we're out of time this week.  I'll talk to everyone next week.
17:56:52 <tdawson> Bye everyone.
17:57:05 <tdawson> #endmeeting