21:01:22 #startmeeting 21:01:23 dabo: depends on the level of amusement, but in essence...yes. 21:01:23 Meeting started Tue Apr 12 21:01:22 2011 UTC. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:24 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:01:35 Welcome everyone to our weekly OpenStack team meeting... 21:01:42 Today's agenda: 21:01:46 #link http://wiki.openstack.org/Meetings 21:01:58 #topic Current release stage: Release 21:02:14 At the end of the day we are supposed to set RCFreeze and cut a release candidate tarball... 21:02:25 See http://wiki.openstack.org/RCFreeze 21:02:40 With no more merges allowed, we use the last day before release to test the release candidate for blatant regressions. 21:02:55 Let's move directly to the next topic to discuss if a short delay would buy us a better release... 21:03:06 #topic Cactus Release Go/no-go 21:03:25 So there were far too many fixes landing in this post-Gamma timeframe, most of them warranted by their impact... 21:03:41 I think we could use one more day of testing and rc-bugfixing. 21:03:50 Easily. 21:03:56 So my proposal would be to delay the release by one day. 21:04:00 +1 21:04:10 RCFreeze EOD Wed, RC cut Thursday, Release push Friday 21:04:30 I think if we delay by more than a day, we'll start introducing large bugfixes again, with their potential for regression. We'll also open Diablo one week later... 21:04:31 Would the delay be for fixing regressions only? 21:04:54 blamar: the delay would be to fix a list we come up with on the next topic of this meeting. 21:04:57 ttx: do we really want to push a release on Friday ... there will be very little support on IRC during the weekend if we get a show stopper 21:05:03 are there any unfixed bugs still remaining? 21:05:16 vishy: yes, see next topic 21:05:35 kpepple: in fact you cut a release candidate on the Thursday. 21:05:51 kpepple: on Friday it's just about pushing that RC through as the gold release 21:06:26 kpepple: if the RC is broken, we might need more time to fix and cut a new RC. 21:06:41 kpepple: in which case we might delay to next week, but I'd rather not :) 21:07:08 I think if we delay by more than a day, we'll start introducing large bugfixes again, with their potential for regression. We'll also open Diablo one week later... 21:07:29 jbryce, sparkycollier, spectorclan: I hope that won't screw up any PR plan ? 21:07:31 * jaypipes votes for one day delay only. 21:07:37 what are we gaining by a day? 21:07:43 ttx: we should be fine, imho 21:07:46 vishy: a few extra fixes 21:07:51 ttx: we won't really find out if RC is broken until the masses get their hands on it ... but i understand your relunctance 21:08:09 vishy: maybe we should discuss the list first 21:08:21 vishy: and then revisit if a one-day delay is useful. 21:08:21 ttx: ok 21:08:32 #topic RC bugs 21:08:33 ttx: it won't be a problem if we know ahead of time 21:08:43 At this point we have the following RC bugs left: 21:08:49 Nova: https://bugs.launchpad.net/bugs/752735 (Volumes aren't zeroed out) 21:08:50 Launchpad bug 752735 in nova "Volumes aren't zeroed out" [High,In progress] 21:09:00 Glance: https://bugs.launchpad.net/bugs/758761 21:09:02 Launchpad bug 758761 in glance "logging in Swift backend store does not work" [Medium,In progress] 21:09:07 Glance: https://bugs.launchpad.net/bugs/756886 21:09:08 Launchpad bug 756886 in glance "is_public must be "True" - mismatch with docs" [Low,In progress] 21:09:11 jbryce: agreed, if we know new plan today it's fine 21:09:14 Glance: https://bugs.launchpad.net/bugs/758762 21:09:15 * vishy got smoketests passing again so I'm happy 21:09:15 Launchpad bug 758762 in glance "Need for more user-friendly docs" [Wishlist,In progress] 21:09:19 Swift is good to go so far. 21:09:39 Those 4 could go in today. 21:09:48 But I have other candidates on my watchlist: 21:09:53 Bug 759053 21:09:54 Launchpad bug 759053 in nova "Running an instance using an image name returns NotFound" [High,In progress] https://launchpad.net/bugs/759053 21:10:06 vishy: that's a regression, right ? 21:10:12 yes 21:10:34 and it will make a lot of our installation instructions not work if we don't fix it 21:10:41 ok, so I think it's a good RC candidate ? 21:10:50 ttx, the volumes being zeroed needs fixing - fyi 21:11:11 * ttx add 759053 to the list 21:11:19 Bug 758404 21:11:21 Launchpad bug 758404 in nova "euca-attach-volume fails when using RBD driver" [High,In progress] https://launchpad.net/bugs/758404 21:11:39 If I read this one correctly, RBD volumes are unusable 21:12:31 the fix is rather small and only affects RBD, which is broken without... 21:12:37 opinions ? 21:12:46 should it make it to "the list" ? 21:12:49 that one seems low impact i'd put it in 21:13:04 agreed 21:13:12 Bug 758810 21:13:13 Launchpad bug 758810 in nova "VMwareAPIConnection class should inherit ComputeDriver class" [High,Confirmed] https://launchpad.net/bugs/758810 21:13:39 I'm a bit unsure of the impact, but I suppose it renders VMWare unusable ? 21:13:55 id really like this one to make it as well: https://code.launchpad.net/~vishvananda/nova/has_role_cache/+merge/56449 21:14:00 I also suppose so, thought I haven't had a chance to speak with Steesh 21:14:24 vishy: just a sec 21:15:11 Looking at the patch, I'd say bug 758810 should be in list 21:15:11 Launchpad bug 758810 in nova "VMwareAPIConnection class should inherit ComputeDriver class" [High,Confirmed] https://launchpad.net/bugs/758810 21:15:26 any +1 / -1 ? 21:15:32 ttx: when doing a fix for another bug, I discovered this: https://bugs.launchpad.net/glance/+bug/759012. I believe it should also be in the RC for Glance. It's a pain in the butt for users if it isn't... 21:15:33 +1 21:15:34 Launchpad bug 759012 in glance "verbose and debug settings no effect on logging" [Critical,In progress] 21:15:46 vishy: You want to add a patch that adds a feature, and changes flags? Really? Now? 21:15:46 jaypipes: beh 21:15:58 +1 21:16:11 please only discuss one at a time 21:16:13 jaypipes: do we have the max image size bug fixed? 21:16:22 I'll come back to vish and Jay's. 21:16:31 k 21:16:32 ok, adding the VMWare one 21:16:38 soren: actually I wanted it a couple of weeks ago, but no one was reviewing it :( 21:16:42 bug 742256 21:16:44 Launchpad bug 742256 in nova "nova-compute should logout an iSCSI target on migration failure" [Medium,In progress] https://launchpad.net/bugs/742256 21:16:53 Not completely convinced by this one 21:17:08 since it's an error case, and we are already struggling with normal cases 21:17:10 vishy: That's a separate (but real!) problem. 21:17:23 bug 750544 21:17:24 vishy: max-image-size issue is fixed 21:17:24 Launchpad bug 750544 in nova "Snapshotting VM instance failing for KVM" [High,Fix committed] https://launchpad.net/bugs/750544 21:17:34 vishy: that image size needs 2 more core reviews. 21:17:35 "Fix committed"? 21:17:38 fix commtted ? 21:17:41 wtf 21:18:01 ah, EBADSTATUS 21:18:05 sirp_, vishy: https://code.launchpad.net/~johannes.erdfelt/nova/bug723996/+merge/57367 21:18:12 it's proposed for merging 21:18:34 jaypipes: oh i was thinking of the other issue, the 32-bit size column 21:18:45 nm 21:18:50 if there's any other feedback for that fix, let me know 21:18:53 any opinion on the snapshotting one ? I think this is mostly untested 21:19:15 wow I hadn't seen the snapshotting bug 21:19:30 ttx: Very much so, I have a branch linked, but not confident in it for something to put in the last day. I'm sad to not include it though. 21:19:30 I'm not super comfortable with it. 21:19:47 I'm ok to set it as a known issue 21:19:50 is that doing volume snapshots, too? 21:19:53 Its current breakage is well understood: Attempts just utterly fail. 21:20:05 jaypipes: i was referring to the one in glance that put size as 32 bit number 21:20:13 soren: Have you tried the patches? 21:20:13 ok, -1 21:20:20 If we fix the call to qemu-img, I don't know if it works, and if not, how it fails. I say leave it. 21:20:41 soren: but what is the drawback of fixing that call? 21:20:47 soren: it is already failing 21:20:54 vishy: I don't know. 21:21:00 vishy: And that is precisely the poiint. 21:21:03 vishy: it may work, but you don't document how unsure we are that it does 21:21:14 vishy: that is in already, yes. 21:21:18 vishy: Maybe it overwrites other snapshots. Or kills your kitten. 21:21:19 I don't know. 21:21:29 Right now, it jsut doesn't work. I know how to explain that to people. 21:21:41 And people understand that. 21:21:49 soren: understood. Just marking off the list of things that we need to fix in our deploy branch :) 21:21:57 next: https://code.launchpad.net/~vishvananda/nova/has_role_cache/+merge/56449 21:22:17 vishy: I'm not comfortable with this at this point. 21:22:25 vishy, soren: is 750544 a Glance bug? I can't really tell. Looks like the snapshotting code for libvirt isn't calling glance-upload properly? 21:22:25 checking now, I'm seeing that we *are* doing snapshots... in this case, I'm going to have to more strongly oppose to the wipe-on-delete patch. 21:22:25 ttx: ok 21:22:41 bug 759012 21:22:42 Launchpad bug 759012 in glance "verbose and debug settings no effect on logging" [Critical,In progress] https://launchpad.net/bugs/759012 21:22:45 jaypipes: no i don't think it is a glance bug 21:23:13 jaypipes: I'm ok with it, if you think you can get all of them in in time... 21:23:14 jaypipes: There might be a glance component to the bug, too, but first of all, the call to qemu-img is being passed as a string with spaces and stuff, and not as an argv. 21:23:25 ttx: I have the fix for that done. 21:23:42 jaypipes: ..and that's my problem with it. I don't know if the breakage stops there. 21:23:48 soren: k 21:23:51 ...but I think I've made my point by now :) 21:23:57 heh 21:23:58 jaypipes: added 21:24:09 any other candidates ? 21:24:11 ttx: cheers. Jordan will be happy. 21:24:12 Current lists: 21:24:19 https://launchpad.net/nova/+milestone/cactus-rc 21:24:24 https://launchpad.net/glance/+milestone/cactus-rc 21:24:30 https://launchpad.net/swift/+milestone/1.3-rc (empty) 21:24:51 hrm 21:24:55 I missed the vmware one 21:25:47 That's 4 for Nova, 4 for Glance 21:25:52 warranting one extra day. 21:26:07 *nod* 21:26:14 ttx: k. so when is the new proposed tarball cut time? 21:26:21 We'll need someone to convert the fix for bug 758810 into a branch merge proposal 21:26:22 Launchpad bug 758810 in nova "VMwareAPIConnection class should inherit ComputeDriver class" [High,Confirmed] https://launchpad.net/bugs/758810 21:26:31 (patch is attached to the bug) 21:26:52 We freeze tomorrow EOD, I cut the tarball early Thursday. 21:27:05 I'll drop an email to Sateesh for the Branch merge proposal. If he can't I'll do it. 21:27:11 then unless we discover it's eating your data and sleeps with your wife, it's released on Friday. 21:27:51 the hard-drinkin' cactus... 21:28:01 Last comments ? 21:28:54 and then we open the diablo floodgates on Friday. 21:29:42 we could release like this every week or two :) 21:29:51 Or every day! 21:29:55 jbryce, spectorclan, sparkycollier: so release is early Friday morning 21:30:01 ttx: cool 21:30:34 #topic Open discussion 21:31:13 we need nova-adminclient in the ppas 21:31:15 any one else have doc fixes? 21:31:17 -core teams: At this point please only set to 'accepted' the merge proposals that are on the RC lists 21:32:13 ttx: can we put in doc changes? because if not, I'd like this to go in as well: https://code.launchpad.net/~vishvananda/nova/cloudpipe-docs/+merge/57241 21:32:27 * ttx looks 21:32:33 ttx: I think doc fixes are fine, no? 21:32:39 pure doc fixes... 21:32:44 we've done pure rst edits in the past 21:32:51 I think so 21:33:05 well, they can regress too, like borking the doc generation and not ending in tarball 21:33:17 but we can certainly add them 21:33:44 ttx: true, that doc gen bug is problematic 21:33:57 vishy: i'll create a "cloud pipes is heavily underdocumented" bug and rc it, so that it appears on the list 21:34:17 (on http://wiki.openstack.org/reviewslist/) 21:34:47 cool thanks 21:35:57 done 21:36:37 Last remark: we'll work on release notes at: 21:36:59 http://wiki.openstack.org/ReleaseNotes/Cactus 21:37:21 That includes a small feature list and known bugs 21:37:30 Please feel free to contribute :) 21:37:41 * ttx will work on that tomorrow. 21:37:55 is there a way to make sure we don't overwrite other people's changes when working on the wiki? 21:38:12 tr3buchet: if there was, wikis wouldn't be so fun. 21:38:21 * annegentle will also work on release notes tomorrow 21:38:24 haha, i'm good with that 21:39:07 anything else ? 21:40:20 tr3buchet: heh. MoinMoin does warn you when someone else has the page opened for edits. Just edit quickly :) 21:40:30 ok then. 21:40:35 #endmeeting