13:00:21 <mburns> #startmeeting oVirt Node Weekly Sync 13:00:21 <ovirtbot> Meeting started Tue Mar 27 13:00:21 2012 UTC. The chair is mburns. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:21 <ovirtbot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:31 <mburns> #chair mburns jboggs pmyers fabiand 13:00:31 <ovirtbot> Current chairs: fabiand jboggs mburns pmyers 13:00:37 * pmyers here 13:00:38 <mburns> #topic Roll Call and Agenda 13:00:41 * jboggs here 13:00:46 * fabiand here 13:00:49 <mburns> 1. Action Item Review 13:00:54 <mburns> 2. Release Status 13:00:59 <mburns> 3. Stateless status and update 13:01:04 <mburns> 4. CIM status and update 13:01:10 <mburns> 5. Other topics 13:01:30 <mburns> anyone have other topics? 13:01:46 <mburns> #topic action item review 13:01:58 <mburns> #link http://ovirt.org/meetings/ovirt/2012/ovirt.2012-03-20-13.02.html 13:02:23 <mburns> #info systemd patches posted 13:02:37 <mburns> #info systemd patches reviewed and merged 13:02:53 <mburns> #info bugs updated for 2.3.0 release 13:03:31 <mburns> #info mburns still hasn't looking into stateless yet 13:03:38 <mburns> but we'll discuss that later 13:03:45 <mburns> #topic release status 13:04:02 <mburns> #info oVirt Node 2.3.0 Released 2012-03-26 13:04:26 <mburns> #link http://www.ovirt.org/wiki/Node_Backlog 13:04:48 <mburns> i updated the 2.3.0 link there, added a 2.5.0 link and an untargeted link 13:05:31 <mburns> oVirt Project release is targeting May for their next release 13:05:59 <mburns> we should target 2.4.0 for end of April, then do 2.4.1, 2.4.2 as needed during that next month 13:06:07 <mburns> thoughts? 13:06:22 <jboggs> sounds ok to me 13:06:32 <fabiand> to me too 13:06:47 <ovirtbot> 14[[07Node Backlog14]]4 !10 02http://www.ovirt.org/w/index.php?diff=2961&oldid=2960&rcid=3045 5* 03Mburns 5* (+7) 10 13:06:56 <mburns> ok, updated backlog page with that date 13:07:12 <mburns> #agreed 2.4.0 targeted for 2012/04/30 13:07:50 <mburns> given that, we need to review the bug list for 2.4.0 and move out bugs that we can't get done 13:08:19 <mburns> jboggs: fabiand: can you see what's assigned to you and what you think you can reasonably get done for that date? 13:09:02 <mburns> jboggs: fabiand: also taking into account other responsibilities you have beyond just ovirt-node 13:09:34 <jboggs> mburns, mine are obscure ones that are probably going to be pushed anyways, can possibly take on some you guys want to push 13:09:35 <mburns> #action jboggs fabiand mburns to review bugs assigned and move to 2.5.0 where appropriate 13:10:00 <mburns> jboggs: i expect most of your time will be spent on plugin development 13:10:13 <mburns> i'd like to get that and stateless in for 2.4.0 13:10:24 <jboggs> ok that works 13:10:25 <fabiand> mburns, we'll have to talk about where else I could be working on 13:10:39 <pmyers> network manager conversion? 13:10:42 <pmyers> that's probably important 13:10:42 <fabiand> :) 13:10:43 <ovirtbot> 14[[07Vdsm TODO14]]4 !10 02http://www.ovirt.org/w/index.php?diff=2962&oldid=2676&rcid=3046 5* 03Danken 5* (+241) 10/* What Can You Do for Vdsm */ 13:10:50 <mburns> fabiand: yes, you've done a good job grabbing bz's 13:11:07 <mburns> fabiand: i'd like to get automation stuff going a bit as well 13:11:07 <fabiand> :) 13:11:13 <fabiand> mburns, definetly! 13:11:19 <fabiand> I'd also like to get some automated testing 13:12:09 <hitvx> suggestion, one should not 'feature freeze' on 2012-04-31 (http://www.ovirt.org/wiki/Second_Release) ;) 13:12:12 <xTs_w> mburns: :) 13:12:19 <xTs_w> mburns: how's the stateless node coming? 13:12:46 <mburns> #action mburns to work with fabiand on mapping out automation plan 13:13:40 <mburns> hitvx: yeah, hard to freeze on a day that doesn't exist 13:14:32 <mburns> we'll target 2012/04/30 for our 2.4.0 and that will allow us to meet the 2012/05/31 release date for the overall project 13:14:40 <mburns> xTs_w: stateless is the next topic 13:16:00 <ovirtbot> 14[[07Second Release14]]4 !10 02http://www.ovirt.org/w/index.php?diff=2963&oldid=2795&rcid=3047 5* 03Mburns 5* (+0) 10/* Timeline */ change feature freeze to 30-April since 31-April doesn't exist 13:16:11 <mburns> ok, anything else for releases? 13:16:34 <mburns> #info oVirt 3.1 release targeted for 2012-31-05 13:16:49 <mburns> #info oVirt 3.1 feature freeze 2012-04-30 13:16:56 <mburns> ok, on to stateless 13:17:01 <mburns> #topic Stateless 13:17:21 <mburns> #link http://www.ovirt.org/wiki/Node_Stateless 13:18:01 <mburns> #action mburns to update wiki page 13:18:10 <mburns> #undo 13:18:10 <ovirtbot> Removing item from minutes: <MeetBot.items.Action object at 0x96af44c> 13:18:15 <mburns> #action mburns to update wiki page for stateless 13:18:44 <mburns> i've been using stateless locally to test images 13:18:52 <mburns> it seems to be fairly stable 13:19:09 <mburns> but we're still limited by the lack of support for stateless nodes by vdsm and engine 13:19:15 <pmyers> xTs_w: speaking of stateless, any chance you can help push that along re: some patches? If we had some more folks helping in the community we could make a lot more progress here :) 13:19:23 <fabiand> mburns, would be nice if you added your notes how to do this to the wiki 13:19:33 <mburns> fabiand: how to run stateless? 13:19:40 <fabiand> mburns, yes 13:19:57 <pmyers> mburns: stateless for vdsm/oVirt Engine is outside the scope of oVirt Node project... we'll do what we can to make Node stateless (as much as possible) and then leave the rest to vdsm team 13:20:05 <xTs_w> pmyers: I'd like to help, but recently had to kill my whole testing equipment, because we needed the parts somewhere else :( 13:20:15 <pmyers> that's a bummer :( 13:20:16 <mburns> fabiand: ok, i'll put up a quick wiki page, but in short, add "stateless" to the kernel command line 13:20:16 <xTs_w> I've currently got no stuff to test 13:20:26 <pmyers> you can do dev work on Node in a VM though 13:20:35 <pmyers> so even if you have a single box with kvm on it, it's good enough :) 13:20:35 <fabiand> mburns, right - I can go trough the scripts to see what's happening :) 13:20:55 <mburns> yes, i do most install testing in vms first, and develop on a vm 13:21:03 <mburns> then move to real hardware once i work out bugs there 13:21:35 * mburns has even been known to run test installs on vms running inside ovirt-engine on ovirt-node 13:21:54 <fabiand> mburns, nesting? 13:21:59 <mburns> overall, most of the basic steps for stateless are done 13:22:02 <mburns> fabiand: mostly 13:22:18 <mburns> you can't run vms on the nested or virtualized ovirt-nodes 13:22:21 <fabiand> mburns, nesting would be nice for testing - but later more on this 13:22:34 <mburns> but you can do most of the rest 13:22:49 <fabiand> at least it would be nice to allow virtualized nodes to be registered with engine - for testing 13:23:25 <mburns> #action mburns to review current stateless state and get a todo list together 13:23:57 <mburns> fabiand: yes, registration should actually work with virtualized nodes 13:24:06 <mburns> they should just come up non-operational 13:24:26 <mburns> ok, anything else to cover on stateless? 13:24:31 <fabiand> right 13:25:05 <mburns> ok, moving on to cim 13:25:10 <mburns> #topic CIM Status 13:25:47 <mburns> #info sblim-sfcb service starts, firewall port is opened, cim password can be set correctly 13:25:58 <mburns> #info remote systems cannot query however 13:26:38 <mburns> DV gave me some pointers on node-devel@ earlier and looking through the logs he suggested, i found that it's failing to load the libcmpiutil.so library 13:26:50 <mburns> this is because we blacklisted that package 13:27:13 * mburns will have a patch soon to un-blacklist it and re-test cim 13:27:45 <pmyers> cool 13:28:18 <mburns> #other topics 13:28:42 <mburns> #action mburns to patch missing libcmpiutil package and retest cim 13:28:46 <mburns> #topic Other Topics 13:29:01 <mburns> Automation 13:29:21 <mburns> fabiand: we can work up a high level plan for automation, then we can send it to node-devel@ 13:29:31 <fabiand> sounds good 13:29:47 <mburns> makes more sense to do that off line to me rather than in a meeting like this 13:30:02 <mburns> Network Manager conversion 13:30:12 <mburns> pmyers: i didn't completely ignore your comment earlier 13:30:37 <mburns> we're going to need to move toward supporting the network manager api in the near future 13:30:39 <fabiand> I'd say we should start this early .. 13:30:51 <mburns> fabiand: yes, i agree 13:31:08 <pmyers> mburns: I know you didn't which is why I didn't poke you a second time :) 13:31:16 <jboggs> using NM api should hopefully clear up alot of the networking scenarios from potential failures 13:31:31 <mburns> it appears that vdsm is going to eventually make NM their default interface for network management, so at the very least we need to start including it 13:31:40 <mburns> jboggs: i hope so as well 13:32:14 <mburns> i think what we'll do is have jboggs focus on the plugins 13:32:28 <mburns> fabiand: on the automation stuff 13:32:40 <mburns> mburns: on the stateless 13:32:43 <mburns> all of us fixing bugs 13:33:05 <mburns> and whoever gets through the big ticket stuff first takes the first pass at designing the NM changes 13:33:50 <fabiand> Maybe the last one getting thourgh should get NM? ;) 13:33:58 <jboggs> haha 13:33:58 <mburns> i'll try to get a wiki feature page up for NM this week 13:34:01 <mburns> fabiand: ;-) 13:34:12 <mburns> #action mburns to post wiki page for NM feature 13:34:32 <mburns> anything else for NM? 13:34:41 <mburns> not sure there is much to talk about at the moment here 13:34:53 <mburns> ok, one other thing from me then 13:34:57 <mburns> #link http://ovirt.org/wiki/Node_Troubleshooting 13:35:05 <mburns> i threw this together yesterday 13:35:18 <mburns> so it's really just a dump of stuff off the top of my head 13:35:26 <mburns> feel free to add anything else you've found 13:35:38 <mburns> or add additional topics 13:36:04 <mburns> or even erase what i wrote as being horribly wrong... 13:36:30 <mburns> that's all i have for today 13:36:32 <fabiand> Don't know if it's relevant for the meeting, but the /data-not-beeing-mounted bug seems quite important to me 13:36:43 <fabiand> because it will make updates difficult, afaiu 13:36:54 <mburns> sure we can cover it briefly 13:37:01 <mburns> #topic mountpoints not getting mounted 13:37:15 <mburns> #link https://bugzilla.redhat.com/show_bug.cgi?id=806349 13:37:38 <mburns> background -- /data is not getting mounted correctly on reboot 13:37:51 <mburns> and i've heard that /var/log isn't either 13:37:55 <fabiand> yep 13:37:59 <fabiand> just Config 13:38:26 <fabiand> thus, logging and data as well as bind-mounting /data/images to the libvirt path doesn't happen 13:38:47 <mburns> fabiand: has investigated and found that this is due to a systemd bug where it recognizes that the unit file that does the /etc/fstab mounting was updated, but doesn't reload 13:39:03 <fabiand> I wouldn't say it's a bug .. 13:39:06 <mburns> he's posted a workaround fix already and filed a systemd bug on this 13:40:00 <fabiand> (swap is also affected) 13:40:38 <fabiand> At least: A couple of paths are not mounted correctly, including the images path and that ain't to good in my eyes 13:41:07 <mburns> /data/images is only a problem if the user wants to upgrade from engine or run a "local" storage domain 13:41:42 <mburns> HostVG-Logging and Swap not working right are larger concerns for me 13:41:53 <fabiand> ah okay - I thought that /data/images is the "main entry point" 13:41:59 <fabiand> (for storing images) 13:42:39 <mburns> i'm pretty sure vdsm manages that in another location, though the actual location escapes me atm 13:43:11 <fabiand> I believe it will be libvirts default location (~ /var/lib/libvirt/images) 13:43:36 <mburns> no, it's something like /engine/data-center or something like that 13:43:44 <fabiand> oh that one. ok 13:43:54 <mburns> they do whatever mounting they need there 13:44:12 <mburns> anything else to discuss on this? 13:44:18 <mburns> i think it's well in hand currently 13:44:29 <mburns> we have a workaround and bug filed for the original problem 13:44:35 <mburns> and a separate bug to revert the fix 13:44:44 <mburns> err...revert the workaround 13:45:10 <mburns> ok, thenk 13:45:14 <mburns> #topic Other Topics 13:45:23 <mburns> anything else to bring up today? 13:45:50 <mburns> going once... 13:46:05 <mburns> twice.. 13:46:16 <mburns> gone 13:46:22 <mburns> thanks everyone 13:46:25 <mburns> #endmeeting