14:09:50 #startmeeting 14:09:50 Meeting started Mon Jul 14 14:09:50 2014 UTC. The chair is dcaro. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:09:50 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:09:55 * rbarry here 14:10:06 hello! 14:10:07 #chairs eedri obasan rbarry 14:10:08 * obasan here 14:10:47 * fabiand is here 14:11:09 #chairs fabiand 14:11:28 #chair fabiand 14:11:28 Current chairs: dcaro fabiand 14:11:32 #chair eedri obasan rbarry 14:11:32 Current chairs: dcaro eedri fabiand obasan rbarry 14:12:23 not too much assistance from outside :/, let's get started anyhow 14:12:41 #topic Hosting 14:13:11 #info phx lab still waiting for final networking setup (cabling and swithces) 14:14:25 there have been no issues with storage on linode lately right? After eyal's cleanup? 14:14:45 dcaro, not that i know off, though jenkins master has out of space this morning 14:15:27 eedri: yep, it's where all the artifcts are stored (and allthe jobs logs), lately we are starting to store isos and other big artifacts... 14:15:52 eedri: was it any particular process taking the space? 14:16:04 dcaro: here 14:16:53 ewoud, publishers were talkign lots of space 14:16:57 dcaro, ^ 14:17:04 dcaro, i didn't really solve the problem, just mitigated it for now 14:17:13 dcaro, a deeper look and adjustments is needed 14:17:22 dcaro, also, 3.3 jobs i changed some of the history there 14:17:36 dcaro, looking forward we'll migrate it to a vm on phx1 14:17:37 2 14:17:52 #action review publishers to leverage disk usage 14:18:23 eedri: but the publishers do not archive any artifacts right? 14:19:07 dcaro, thier workspace took space 14:19:12 dcaro, master tool 4.5 gb 14:21:36 eedri: mmm.... maybe we are fingerprinting artifacts or something :/, will have to look 14:22:03 ok, any other issue with hosting? 14:22:26 dcaro, any update on os1 with centos 7 14:22:35 dcaro, now that centos7 is publically availiable 14:23:01 eedri: not really, but we are getting short in space there also 14:23:18 dcaro, really? 14:23:24 dcaro, they have a quaote 14:23:37 eedri: we don't have enough quota to launch a new xlarge instance 14:25:05 (last time I tried at least, I created a small instance instead 14:25:34 eedri: maybe we can review current instances and clean up a bit 14:25:45 dcaro, ok, i guess we should stop relying on it and push forward phx2 14:26:45 eedri: sure 14:28:07 ok, we can move then 14:28:17 #topic Jenkins 14:28:46 eedri: obasan you did a good job creating the 3.5 jobs right? What's still missing? 14:29:20 dcaro, yea, also repo closure was fixed for master & 3.5, though there are some errors still there from time to time 14:29:29 dcaro, still missing aaa projects & hosted engine ha 14:29:47 dcaro, afair.. also vdsm-json-rpc for el6 needs fixing (it had issue with maven + mock..) 14:29:57 eedri: I need help from Jay 14:30:11 eedri: I created the errata and builds, but I can't add them there 14:30:42 eedri: I already sent him an email 14:30:44 msivak, we're in the infra meeting, i will sync with you later 14:31:02 #chair ewoud 14:31:02 Current chairs: dcaro eedri ewoud fabiand obasan rbarry 14:32:02 dcaro, ok, let's continue.. 14:32:03 #action create aaa projects jobs 14:32:16 dcaro, also - we're still missing some jobs on 3.5, for verifying patches 14:32:31 eedri: any specific project? 14:32:33 dcaro, like upgrade db, dao tests and other 14:32:44 dcaro, upgrade db has open trac ticket on it 14:33:13 eedri: ok, we'll review later then 14:36:37 what about the copr jobs? Will we keep pushing them? 14:37:49 dcaro, yea, i think it will make life easier for official releases, but not high priority i guess 14:38:04 ok 14:38:34 #topic puppet/foreman 14:39:49 dcaro, about jenkins, we have failures also on upgrade jobs 3.3/3.4 14:39:55 there was a discussing a couple weeks ago about how to arrange the manifests, anyone did not have time to participate? 14:39:57 dcaro, not sure if these are known bugs or not 14:40:09 #info failures on upgrade hobs 3.3/3.4 14:40:22 #action check if there are open bugs on 3.3/3.4 upgrade failures 14:41:04 dcaro, also vdsm job failed due to new python-pthreading 14:41:14 dcaro, should be OK now, after you added them to the ci-tools repo 14:41:20 eedri: but that's already on ci-tools 14:41:22 yeah 14:41:30 let's check if it did xd 14:42:17 looks ok, master right? 14:45:04 about the puppet manifest layout, I'll try to gather the comments and make an etherpad so we can decide what to do 14:45:22 #action gather puppet layot discussion arguments and create a pad 14:46:11 #topic tickets 14:47:06 let's review as many tickets as we can 14:47:11 with the time left 14:47:22 https://fedorahosted.org/ovirt/ticket/188 - ovirt-appliance can not be build because of unreachable mirror 14:47:29 fabiand: ^ any updates on this? 14:47:38 dcaro, yep 14:47:50 dcaro, you solved most of it, and I played a bit with proxying to improve it. 14:47:54 So we can close it 14:48:01 and thanks a lot dcaro ! 14:48:05 to you, personally! :) 14:48:34 fabiand: good job with the proxy, I'm looking forward to get it set up on all the slaves 14:49:00 yep, I'll write it up and then we can use it elsewhere 14:49:45 https://fedorahosted.org/ovirt/ticket/144 - ovirt-reports_3.4_create-rpms_gerrit needs jasperreports-server >= 5.5.0-6 14:50:21 eedri: sbonazzo this ^ was solved? Partially solved? 14:51:18 dcaro, was the job migrated to mock? 14:52:03 eedri: no 14:52:07 dcaro http://jenkins.ovirt.org/job/ovirt-reports_master_create-rpms_merged/286/console looks ok so maybe it has been fixed 14:52:18 dcaro I don't know what changed 14:52:35 that's why I'm asking, maybe we should close that one and open a new one to migrate the job to mock 14:53:21 dcaro +1 14:53:26 dcaro, sounds right to me, via yaml when it's ready 14:53:41 dcaro, also not critical for sure 14:53:45 okok 14:55:22 https://fedorahosted.org/ovirt/ticket/204 - [jenkins][java] fix issues where JAVA_HOME has different locations on various slaves 14:55:53 I forgot about this one, I have to fix it, was gonna create a non-mvn job for it so we ca use any system installed java instead of having to specify one 14:56:10 I'll accept it 14:56:25 dcaro, i'm not sure it's relvant fpr non-maven jobs 14:56:41 dcaro, it happens if jobs are not set to 'default' java on matrix jobs 14:56:44 dcaro, iirc... 14:56:44 eedri: yep, that's why move it to non mvn 14:56:55 dcaro, way before the mvn section 14:56:59 dcaro, right after the clone 14:57:11 eedri: yes, it's that two combination 14:57:19 dcaro, matrix jobs are "mvn type" 14:57:20 ? 14:57:42 eedri: noo 14:57:45 *no 14:57:52 mm... *nop xd 14:58:00 dcaro, so not sure you're right about the mvn tyhpe 14:58:14 dcaro, i think it's related to the jdk you choose on the job of the job, 14:58:26 eedri: when creating a multiconfig job, you can add a maven build step 14:58:36 dcaro, yea i know 14:58:46 then the java installation becomes relevant 14:58:55 just as if it were a maven type job 14:59:32 dcaro, for example - http://jenkins.ovirt.org/job/ovirt-engine_3.4_find-bugs_merged/configure 14:59:40 dcaro, is this a maven job or freestyle 15:00:44 dcaro, see the JDK option right after the params 15:00:51 eedri: I'd say it's freestyle 15:00:59 dcaro, ok, so that job was failing 15:01:16 dcaro, cause it had some jdk definition choosen which didn't match the jdk path on the slave 15:01:25 eedri: it has configured a specific installation of jdk, not told to use the default one 15:01:29 dcaro, i suspect cause the path on the master is different from the slave 15:01:54 eedri: anyhow, to solve it, just remove the maven build step and change it for a shell script runnin gmaven 15:01:55 dcaro, yea, so i wonder how you overcome an issue where the jdk is on one path on the master and different on the slave 15:02:14 dcaro, if you do that, it won't try to use the "JDK" value from the job? 15:02:28 eedri: no afaik 15:02:31 dcaro, i see 15:02:41 dcaro, ok, so then you have a different problem :) 15:02:46 it uses the system PATH 15:02:47 dcaro, no official maven on el6 15:02:55 dcaro, unless we have a custom rpm for it.. 15:03:10 if mvn is installed system wide, it should not care where it comes from 15:03:36 and we have a custo maven3 rpms for el6 on ci-tools repo iirc 15:04:05 dcaro, ok, so we're good 15:04:50 (got deleted not long ago xd, but we restored it soon after) 15:05:53 okok, so for the record: 15:06:09 i'll write it down in the ticket xd 15:06:51 hehe 15:06:52 dcaro, ok 15:07:47 ok, last ticket 15:07:53 https://fedorahosted.org/ovirt/ticket/140 - [jenkins] ovirt-node jobs issue 15:08:49 fabiand: ^ any insight on i, have you looked? 15:11:59 ok, fabiand is not around xd 15:12:21 okok, so, anything before we end the meeting? 15:13:23 obasan: eedri ewoud misc orc_orc rbarry ? 15:13:41 dcaro, hmmm 15:13:51 Nothing from me 15:14:02 dcaro, fabiand rbarry what about ovirt-node publishers 15:14:17 currently we don't publish ovirt-node via nightlies 15:14:21 eedri, I'm getting squashed between two meetings right now. 15:14:30 fabiand, hehe.. ok, will take this offline 15:14:34 eedri, the biggest issue is that ovirt-node still does not build. 15:14:39 The ovirt-node master rpm job is working, ISOs aren't 15:14:52 fabiand, oh.. that will make adding nightlies a problem then :D 15:14:53 repo problems with the ISO build, I think. But hopefully building this week 15:15:18 sbonazzo, someone talked about improving nightlies usage, either by ovirt-release rpm or documentation 15:15:33 sbonazzo, people might not be aware they need to use other repos like pattenfly..and so on 15:15:50 sbonazzo, i think there is a ticket on it 15:16:40 okok, I'll close the meeting so we can take this offline 15:16:44 dcaro, +1 15:16:44 #endmeeting