14:01:07 #startmeeting oVirt Node Weekly Meeting 14:01:07 Meeting started Tue Jun 4 14:01:07 2013 UTC. The chair is mburns. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:07 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:01:14 #chair jboggs rbarry fabiand 14:01:14 Current chairs: fabiand jboggs mburns rbarry 14:01:17 * fabiand is here 14:01:20 #topic agenda 14:01:27 * rbarry is here 14:01:32 #info patch review 14:01:35 #info feature status 14:02:10 BlairQ: try disabling ovirt-beta and enabling just ovirt-stable 14:03:18 #topic outstanding patches 14:04:17 #info 25 patches outstanding, some with acks that just need dependencies merged or rebasing 14:04:58 * fabiand posted a set of patches to do some package cleanups 14:05:00 #info many patches around cleaning up old code 14:05:06 thanks to mburns and jboggs we could remove some unused files 14:05:22 #info some draft and RFC patches that need updates 14:06:05 #info 6 patches have code review nacks that need action 14:06:48 #action ALL -- review patches, update patches that need updates, nag people for reviews if your patches aren't being reviewed 14:07:01 * jboggs here 14:07:15 * mburns makes a note to start doing nag emails... 14:07:46 anyone have other comments on patch status? 14:08:52 nope 14:08:53 well 14:09:06 I plan to push some patches wrt to the brodgeless mode during the next two days .. 14:09:14 I suppose we'll need to do some testing on them .. 14:09:16 fabiand: yep, that's next topic 14:09:37 #topic Feature review 14:09:52 #info bridgeless network -- patches to be posted in next couple days 14:11:09 fabiand: agreed, we need a *lot* of testing with this feature 14:11:19 as in full end-to-end testing 14:11:24 yes 14:11:38 maybe we can take the opportunity to write some tests like you did with the vdsm plugin .. 14:12:46 fabiand: yes 14:13:04 rbarry: any updates on the password widget? 14:13:19 I'll open a feature page for the bridgeless stuff and prepare a table, 14:13:44 mburns: Patches posted today 14:13:56 fabiand: excellent! 14:14:08 #action fabiand to create bridgeless network feature page and test matrix 14:14:53 fabiand: you should link it to the oVirt 3.3 release page, node 3.0 release page, and engine feature page that it's related to 14:15:04 yep 14:15:13 rbarry: awesome 14:15:22 #info fabiand to link it to the oVirt 3.3 release page, node 3.0 release page, and engine feature page that it's related to 14:15:36 #info patches for password/confirmation widget are coming today 14:17:17 #info tuned stuff is close, updated patches should be available in next couple days 14:17:30 nice 14:17:33 #info log sync should also be updated in that timeframe 14:17:46 #undo 14:17:46 Removing item from minutes: 14:18:07 #info log directory sync from iso to filesystem should also be updated in that timeframe 14:18:31 jvandewege: ok, need to check that... 14:18:40 #info EL6 is close to working 14:18:56 jvandewege: have you ever switched nics on IP base ? I did it (and before) changed mac-address but no network anymore... weird ? 14:19:03 #info EL6 needs a couple outstanding patches merged (dracut, urwid workarounds) 14:19:22 fabiand: how is F19? 14:19:38 mburns, fine the last time I looked! 14:19:49 mburns, I had a couple of patches ... one bigger issue with selinux 14:19:52 but now it looks okay .. 14:19:58 fabiand: excellent 14:19:59 but we need to fix collectd .. 14:20:02 also on F18 .. 14:20:08 fabiand: yes 14:20:12 it really messes up the logs 14:20:15 fabiand: is there a bug? 14:20:20 yep 14:20:43 .bug 908270 14:20:43 fabiand: Error: "bug" is not a valid command. 14:20:45 Yamaksi_: switching nics also changes mac and that means no network when using linux. /etc/udev/xx-network-abc controls that 14:21:24 mburns, and - I didn't test it but biosdevname is also a topic .. 14:21:50 I didn't have time to look at the ... i guess problem ... in it's completeness, but our udev rules wouldn't have any affect anymore 14:22:27 oh, wonderful 14:22:52 fabiand: maybe take that into consideration with your bridgeless network effort? 14:22:56 jvandewege: mhh, ubuntu is better on it 14:23:04 mburns, yes ... 14:23:17 jvandewege: I changed the mac's also 14:23:26 jvandewege: and removed that file and rebooted 14:23:36 mburns, it's mainly investigation for now, and el6 dependent ... iiuic the nic names will be predicatble in f19+ ... without any need for a configuration file .. 14:23:57 Just wanted to give heads up 14:24:07 fabiand: ok 14:24:18 we'll have to watch it in f19 and el6 14:24:26 yep 14:24:41 * mburns commented on the collectd bug 14:24:55 ok, any other features that we need to discuss? 14:25:05 oh, puppet 14:25:34 * fabiand reviewd rbarry s patch. it looks quite good now, found just a couple of small .. glitches 14:25:42 But I haven't tested it yet 14:25:55 ok 14:26:13 jvandewege: okay, something weird in that file, two times eth0 14:26:51 #info puppet patch is close, some comments and small tweaks needed 14:27:05 The patch is much trimmer, specfiles/makefiles merged. fabiand's comments should be addressed today or tomorrow at the latest, depending on how long finishing the password entry stuff takes me 14:28:47 ok, anything else that i'm forgetting? 14:29:29 not from me 14:30:07 ah well 14:30:07 jenkins problems? 14:30:35 fabiand: tbh, i'm just about ready to say forget jenkins.ovirt.org 14:30:45 mburns, why? 14:30:55 and - what alternative to use? 14:31:29 fabiand: we don't have an alternative, but it's just painful that we can only get about 5 runs before the slave breaks... 14:31:47 is there a way that we can use the jenkins virt plugin 14:31:49 and it's not always our fault... 14:31:51 or jeknkins libvirt .. 14:31:55 mburns, yes ... 14:32:11 not familiar with the jenkins virt plugin 14:32:13 with the jenkins libvirt (or virt?) plugin we could launch a vm on each build .. 14:32:20 the same vm .. 14:32:28 at least we don't need to reboot the host .. 14:32:46 fabiand: problem is that we're running on a vm already 14:32:59 argh ... 14:33:07 * fabiand is not happy about the situation 14:33:16 and I'd like to keep the continous builds .. 14:33:27 fabiand: only alternative i see is to run it on the internal systems and publish to ovirt.org somehow 14:33:30 but that's not ideal 14:33:34 no 14:33:37 what about nesting? 14:33:58 didn't someone from infra, was it ewoud?, look at if we could use nesting? 14:34:14 fabiand: oh, wait... 14:34:24 * fabiand freezes .. 14:34:29 fabiand: maybe, if the virt plugin supports it 14:34:35 we could have it launch our own vm 14:34:42 dedicated to our use for building 14:34:46 that could work 14:34:55 but it needs to be able to work with ovirt 14:35:04 mburns, sounds good! 14:35:05 and i'm guessing that's not something it can do 14:35:09 mh .. 14:35:21 cant we use ovirt's api to trigger a vm boot from the job .. 14:35:49 mh 14:35:53 fabiand: it gets ugly, doing that 14:35:58 yep 14:36:08 then we have to do all that setup, ssh to the vm once it's up 14:36:10 but it is already ugly at a different node in the graph 14:36:11 then run everything 14:36:49 mburns, I imagine that we prepare one vm builder and jenkins slave, which is down, then brought up to do the build, then brought down again .. 14:36:50 fabiand: for nested virt: the rackspace servers are installed with fedora, I think kiril, eedri or dcaro was planning on installing ovirt on them 14:37:11 ewoud, awesome! that means nesting can work ... iiuic 14:38:36 fabiand: we also have a pending switch to live-media/oz 14:38:48 that might help this longterm as well... 14:39:14 in the meantime, i think we just have to continue with the occasional reboot of the jenkins host 14:39:36 although maybe we can try to trace what's doing the other stuff that's leaving loop devices around 14:40:06 I once requested a temporary access to the builders to trace it down -.. 14:40:16 but folks were sceptical about this .. 14:40:44 I can understand that, I'd still offer to try to track it down .. if noone else volunteers .. 14:40:57 fabiand: i'll ack access for you 14:42:02 mburns, I'd request the access on the infra ml again .. 14:42:08 fabiand: ok 14:42:18 do so, and i'll ack and argue if needed 14:42:23 ok, anything else? 14:43:46 ok, let's close the meeting then, and try to get these patches reviewed/acked/merged 14:43:49 #endmeeting