14:02:58 #startmeeting oVirt Infra 14:02:58 Meeting started Mon Mar 31 14:02:58 2014 UTC. The chair is knesenko. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:02:58 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:03:08 #chair dcaro ewoud eedri 14:03:08 Current chairs: dcaro eedri ewoud knesenko 14:03:16 bkp: meeting ? 14:03:26 * orc_orc is here but has been travelling 14:03:38 * bkp here 14:05:17 #chair bkp 14:05:17 Current chairs: bkp dcaro eedri ewoud knesenko 14:05:33 Hello all how are you today ? 14:05:46 Pretty well. 14:06:39 #topic Hosting 14:06:54 what's the status with rackspace03 ? 14:07:14 knesenko, last i heard it's been migrated to our vlan and waiting as new hypervisor 14:07:34 knesenko, we can start adding new VMs on it, but we need an easier way to do that 14:07:46 knesenko, so having a pxe server from foreman is the next step 14:08:01 correct 14:08:10 ewoud: any progress on PXe installation ? 14:08:16 did you had time for it ? 14:08:48 fabiand: what to join the infrea meeting ? 14:08:55 Yep! 14:08:56 knesenko: no, been very busy 14:09:00 ewoud: I see ok 14:09:01 knesenko, thanks for the reminder! 14:09:02 :) 14:09:11 #chair fabiand 14:09:11 Current chairs: bkp dcaro eedri ewoud fabiand knesenko 14:09:18 * fabiand is here 14:09:30 dcaro: any chance we can push it this week ? 14:09:35 knesenko: I intend to work on it this evening 14:09:45 ewoud, +1 14:09:49 ewoud: ok thank you very much for it 14:09:54 ewoud++ 14:10:19 #info ewoud will install PXE server on rackspace env this evening 14:10:23 also, this weekend I leave for a week so I really want to complete it this week 14:10:25 anything else on hosting ? 14:10:31 knesenko, yes 14:10:41 eedri: yes please 14:10:47 knesenko, i added 2 new slaves for jenkins running on OpenStack1 public cloud 14:11:04 knesenko, we got access to use it, and dcaro is currently tesitng it to see any issues/performance 14:11:33 eedri: I heard its very slow 14:11:34 eedri: yep! And I'll add a f20 one too, for now, for vered to test the storage functional tests 14:11:34 right ? 14:11:35 so if all goes well, we might be able to add more slaves from there, both slaves are puppetized and accessible to all infra memebers 14:11:52 knesenko, not very, it's 2/3 the speed of rackspace from what i saw 14:12:04 knesenko, for e.g 30 min on unit tests instead of 20 14:12:08 something like that 14:12:10 eedri: so we can run ligth jobs on it 14:12:14 pep8 etc 14:12:30 knesenko, either that or exotic os, like rhel7/ubunto/gentoo etc.. 14:12:37 knesenko, for testing/pakcging 14:12:51 eedri: +1 agreee 14:13:29 knesenko, another update is that servers from the new lab has arrived 14:13:37 knesenko, we should get an update soon on ETA to start using them.. 14:14:04 #info new HW for infra arrivied 14:15:13 #topic Foreman and Puppet 14:15:31 I assume will need to discuss pep8 installation on slaces 14:15:34 slves 14:15:40 on how are we going to do that 14:15:47 we need to new version to be installed right ? 14:15:49 1.5.1 ? 14:15:54 I propose creating a package and pushing it with ci-tools repo 14:16:44 dcaro: should be easy , right ? 14:18:22 dcaro: / 14:18:23 ? 14:19:13 ok nm 14:19:19 lets continue 14:19:49 #action dcaro or knesenko create a pkg for pep8 and deploy it on ci-tools repo 14:19:56 anything else on puppet foreman ? 14:20:06 knesenko, yea 14:20:08 knesenko, sort of 14:20:49 knesenko, nm, i'll mention it in the tickets topics 14:20:59 eedri: ok 14:21:05 #topic Jenkins 14:21:59 dcaro: eedri ? 14:22:00 dcaro, any updated? 14:22:36 knesenko, i think repo_closure jobs are waiting for fabian_a fix.. 14:22:41 knesenko, should be in soon... 14:22:45 fabiand: hey there :) :) :) 14:23:00 :) 14:23:01 fabian_a, didn't you wanted to talk about ovirt-node functional testing ? 14:23:02 hey hey 14:23:11 upgrade job is getting fixed, it has some issues with the yum groups and the repos when upgrading, I'm on it 14:23:17 for the repoclosure jobs ... yes. working on them 14:23:39 eedri, and for the node functional tests ... yes, I would like to get some speed onto that, tho I don't have something righ tnow. 14:23:54 What is important to use i sthat we've got some place somewhere were we can spawn VMs 14:23:55 :) 14:24:01 fabian_a, we do :) 14:24:14 fabian_a, just ping us when you're ready to start working on that 14:24:29 eedri, leave fabina_a alone ;) 14:24:38 eedri, ack, we will do! 14:24:38 fabian_a, sorry :) 14:24:42 ;) 14:24:50 hehe 14:24:54 dcaro, any more updates? 14:25:02 hahahaha 14:25:20 os slaves look promising, but they lack the last push (troubleshooting the connectivity to jboss mvn repos) 14:26:23 *os1 (openstack public cloud) 14:26:40 dcaro, anything you think we can do on the short term to improve infra/jenkins master speed? 14:26:48 when I log on ovirt reports since webadmin ovirt portal I have error java.lang.IllegalArgumentException: An id is required to lookup a FlowDefinition 14:27:26 eedri: split it into two 14:27:28 dcaro, also, i think another function tester requested ssh access to slave 14:27:45 dcaro, 2 masters? 14:27:50 eedri: yes, vered, I'm preparing an os1 f20 slave for him 14:27:58 dcaro, it's she :) 14:28:05 eedri: yes, one with gerrit triggered patches and another with packaging 14:28:28 dcaro, hmm.. ok, let's see if we can run 2 HA vms on the new lab 14:28:31 vered: sorry she :*| 14:28:37 dcaro: :) 14:28:38 dcaro, instead of the the current bare metal 14:30:46 anything else on jenkins? 14:30:49 ah yes 14:30:56 we have some issues with the API 14:30:59 dcaro: right ? 14:31:12 * fabiand would also like ot have slaves access to debug all sorts of Node problems .. 14:31:28 we tried to download artifacts via the API and we had some issues ... 14:31:30 knesenko, What's new in 1.556 (2014/03/23) 14:31:30 Access through API token and SSH key login now fully retains group memberships. (issue 20064) 14:31:37 knesenko, maybe related? 14:31:43 so for now we applied a workaround 14:31:44 dcaro, we can schdule an upgrade 14:31:52 eedri: +1 for upgrade 14:31:58 and see if it works 14:32:07 eedri: knesenko the issue was related to timestamps and non-ssh access 14:32:21 but yes, we might want tot upgrade 14:32:32 Slave started from Java Web Start can now install itself as a systemd service.' 14:32:33 nice 14:32:42 form what I read I its related to git plugin plugin 14:32:44 no need maybe for a puppet class :) 14:33:49 \o/ 14:34:10 paz solved it in a very low level... xd 14:34:59 yea.. 14:35:47 knesenko, there is another updates on repo 14:35:54 knesenko, but that's on repo mgmt topic 14:36:08 eedri: new topic... 14:36:09 ok 14:36:10 knesenko, or builds.. 14:36:16 lets move 14:36:28 #topic Build/Releases/Repo Management 14:36:38 so I have few updates here 14:36:57 if you don't know, me and dcaro are managing resources.ovirt.org repos 14:37:06 and we are release engineers for oVirt 14:37:34 I think we did a great job there .... but we have more in our queue 14:37:38 so updates 14:37:55 #info new layout was applied http://plain.resources.ovirt.org/pub/ 14:37:59 |o/ 14:38:03 +1 14:38:13 #info next z-stream release will be published under pub/ 14:38:19 dcaro, thats playing tennis with one hand? 14:38:19 there are a couple of repos that can be used with both layouts, to ease the transition 14:38:28 #info /releases directory will be obsoleted 14:38:33 * dcaro is about to do an ace 14:38:57 we have automated a lot of process like publishing nightly repos 14:39:12 now we have nightly for 3.3, 3.4 and 3.5 14:39:23 #info we have nightly for 3.3, 3.4 and 3.5 14:39:51 knesenko, dcaro and bkp are working on mirroing those repos to various universities 14:39:52 also we are working on signing RPMs for the official releases 14:40:03 I hope it will be ready for the next zstream release 14:40:33 #info bkp dcaro are working on mirroring repos ti various universities 14:40:38 dcaro: bkp +1 for that 14:41:03 knesenko: dcaro is doing all the heavy lifting 14:42:00 thanks 14:42:03 really 14:42:16 ok what else do we have here 14:42:17 > 14:42:18 ? 14:42:41 trac 14:42:42 tickets 14:43:13 #topic Review Tickets 14:43:16 oh ... 14:43:20 lets do it ... 14:43:36 eedri: would you like to lead this topic ? 14:43:45 knesenko, sure 14:44:05 1st, i'd like to propose something we did with other ovirt project 14:44:06 s 14:44:31 i think we might find help on the community form people wanting to pitch in and assist in various tickets 14:44:41 like adding a puppet class or writing a script 14:44:54 we should write a wiki on ovirt.org on how to help or a landing page 14:45:17 and maybe on occasion ask community if anyone is volenteering to assist in a certain task 14:45:42 it can be communicated to users & devel lists 14:46:23 eedri: the only problem we have is permissions 14:46:30 we have no level for permissions 14:46:37 knesenko, true, but some tasks doesn't require any permisisons 14:46:38 we have only sudo ... and non sudo 14:46:40 knesenko, like sending a patch 14:46:45 eedri: ok 14:46:47 knesenko, you don't need access to anything 14:46:55 knesenko, infra member can assist in verifying 14:46:57 knesenko, if needed 14:47:17 yeah 14:47:18 ok 14:47:33 bkp, do you know who can ask for admin rights for the fedora hosted trac instance? 14:48:01 eedri Not sure. 14:48:07 Lemme check 14:48:16 bkp, also - is this ticket going to solved this week? 14:48:17 https://fedorahosted.org/ovirt/ticket/10 14:48:22 migrating arch list 14:48:39 Yep. In fact, today, I believe 14:48:45 bkp, +1 14:49:33 https://fedorahosted.org/ovirt/ticket/32 14:49:44 enable snytax highliting in media wiki 14:50:01 bkp, dneary what is the process of upgrading media-wiki? 14:50:08 is it done by openshift? or us 14:50:19 eedri, We need to do it 14:50:19 eedri: bkp we also need to apply some security patches right ? 14:50:30 The process is: 14:50:31 dneary, do we have any plans for it? 14:50:37 dneary, other reasons for upgrade.. 14:50:39 eedri: We are truing to figure that out now. There was a security patch issue that came up this weekend, too 14:50:47 *trying 14:50:51 The procedure is: 14:51:00 * Download new MediaWiki 14:51:13 * Over-write original files in Quickstart 14:51:50 (* Deploy to staging, run DB update through the interface, test that it works OK) 14:52:10 * Deploy to prod, run DB updating script through the interface, verift that the wiki is back and all is OK 14:52:35 In brackets because to do that, you need a staging instance 14:52:51 dneary, can we get one from openshift? 14:55:22 dneary, ok, who is going to do the upgrade? 14:56:50 alinefm, ping 14:57:21 hi, eedri 14:57:32 alinefm, hey, how are you? 14:57:42 eedri, I'm afraid I don't know 14:57:50 eedri, good and you? 14:58:14 alinefm, any update on the ibm servers? 14:58:21 eedri, bkp is looking for someone with the skills & comfort level to take it on, I think 14:58:29 dneary, ok 14:58:34 knesenko, can you add it as action item? 14:58:34 eedri: Yes, that 14:58:52 knesenko, planning to upgrade media wiki, still working out the details 14:58:53 eedri: upgrading the media wiki ? 14:58:59 eedri: I found the person who can grant those admin rights? 14:59:23 So I will get them connected to the Infra team. 14:59:25 bkp, i'm not too familiar with media wiki at all, but i assume the credentials for it are i the same place 14:59:29 eedri, well, we finished the machine setup in a lab last week 14:59:43 eedri, now I am working with the legal process 14:59:55 alinefm, nice.. so you'll let us know when can we connect it to jenkins? 14:59:57 eedri Admin rights for the trac instance. 14:59:57 need some contracts to allow the machines be accessible outside IBM network 15:00:06 bkp, ahh.. nice, who is it 15:00:13 Karsten Wade. 15:00:20 (quaid) 15:00:22 quaid! 15:00:24 eedri, sure - I hope get all done in more 2 weeks 15:00:28 yea... figured that 15:00:34 I will send a mail to mail list when it happens 15:00:34 alinefm, btw, you don't nessasarily need that 15:00:45 eedri: Credentials are in same place, yes 15:00:53 alinefm, we have jenkins slaves running in a dmz in tlv, without acccess fromt he outside 15:01:20 alinefm, we're using jnlp connections from the slaves to the jenkins instead (headless mode) 15:01:22 eedri, really? but how do you connect to the slave? 15:01:27 hmm... 15:01:40 alinefm: slave connects to jenkins 15:01:45 and not jenkins to slave 15:02:02 alinefm: so your slaves needs access to the world, that's all 15:02:04 eedri, knesenko, it can helps the process 15:02:33 alinefm, https://wiki.jenkins-ci.org/display/JENKINS/Distributed+builds 15:02:36 eedri: I am pretty it is ... better is to use ssh :) 15:02:51 eedguys we are out of time 15:03:04 any urgent issues to discuss ? 15:03:11 Nope 15:03:16 knesenko, no, we can continue next time. 15:03:32 k 15:03:32 but i suggest going over the tickets and seeing if anything can be requested from the community 15:03:35 thanks 15:03:40 #endmeeting