[00:06] wgrant, mumble? [00:08] Is everyone free to use the LP logo to denote link to Lp profile? [00:09] wgrant: try this please: [00:09] sudo rm -rf /var/lib/rabbitmq/mnesia/rabbit/* [00:09] then start rabbit [00:09] ah, found the license. Ignore me. [00:22] wgrant: hi [00:22] wgrant: have you tried that rabbit fix? [00:22] lifeless: On a call right now. Will test soon. [00:22] thanks [00:31] lifeless: can't move to London: ash cloud. Plus, not enough earthquakes... [01:00] Project devel build #832: STILL FAILING in 5 hr 38 min: https://lpci.wedontsleep.org/job/devel/832/ [01:10] :q [01:11] lifeless: I'm using avahi for now, but any idea how to get the IP address of a guest? [01:16] cat /var/lib/lxc/$host/var/lib/dhcp3/dhchp3.eth0.leases [01:16] or check the libvirt dhcp server log/status [01:17] I was hoping I could inspect the interfaces somehow. [01:18] ask hallyn in #ubuntu-server? [01:35] Hmm. [01:35] Why does avahi get stuck registering on first startup :( [01:43] is there a tag for new person picker bugs? [01:45] person-picker [01:46] mwhudson: But if you file bugs, wgrant will frown at you. [01:46] ah i think it was just https://bugs.launchpad.net/launchpad/+bug/800654 [01:46] <_mup_> Bug #800654: Assign Me link on person picker visible when it should be hidden

< https://launchpad.net/bugs/800654 > [01:46] ... maybe [01:47] RARGH, why does apt-mirror feel it needs to download most of the archive? [01:47] i guess i should get devel running and see if it happens there [01:47] mwhudson: Or qastaging. [01:48] wgrant: good point [01:49] ah bug still present there [01:49] the person picker on $team/+addmember has the "assign me" stuff [01:49] wgrant: does this sound like a bug you know about? [01:49] Is that a problem? [01:49] Apart from the wording? [01:49] wgrant: yes, it's the "add member" picker [01:50] wgrant: neither "remove assignee" or "assign me" make sense [01:50] at all [01:50] "Assign me" does, except it should be "Pick me" or so. [01:50] "remove assignee" does not. [01:50] well maybe [01:50] it's not a /common/ use case for a team [01:50] wallyworld_: ^^ [01:51] i guess if you're an owner you might still want to add yourself [01:51] the assign me link has been on the picker for a while. the above bug was to hide it when not needed [01:52] what's the issue with the links? i'm not quite across what is being said above [01:53] "A while" is about a week, right? [01:53] no, longer [01:53] months [01:53] No... [01:53] the last epic iirc [01:53] Only since jcsackett refactored it a couple of weeks ago? [01:53] Before then it was only on the bugtask assignee picker. [01:53] AFAIK [01:53] no, the refactoring was to make the code which provided the links better and consistent [01:54] across the picker refactoring [01:54] but the functionality was added last epic [01:55] oh, only the bugtask picker [01:55] perhaps that's correct [01:55] yes, i think so [01:56] so perhaps we should plug the "leak" [01:56] only the bugtask picker needs "Assign Me" [01:57] others may want "Pick Me" as mentioned above, but do we really need that? [01:57] http://people.linaro.org/~mwh/silly.png [01:57] my fix for the above bug was for the bugtask case - do not show "Assign Me" if "me" was already assigned to the bug [01:57] wallyworld_: should i file a bug? [01:57] wallyworld_: yeah, i see that now (about the bug i linked to) [01:58] https://bugs.launchpad.net/launchpad/+bug/787595 [01:58] <_mup_> Bug #787595: person picker could have a link to choose myself when I am a valid choice

< https://launchpad.net/bugs/787595 > [01:58] It's new. [01:58] mwhudson: yes please, i think it's bad that way it is atm [01:58] mwhudson: perhaps it's just the wording that is bad [01:59] wallyworld_: i beg to differ in this case [01:59] wallyworld_: particular when i'm already in the team [02:00] fair point [02:00] could you please include the screenshot in the bug [02:00] yay double filing [02:01] that will help reinforce the need for a fix in case whoever reads the bug report can't immediately see the issue [02:01] wallyworld_: https://bugs.launchpad.net/launchpad/+bug/801388 [02:01] <_mup_> Bug #801388: some person pickers show "assign me"/"remove assignee" when that makes no sense < https://launchpad.net/bugs/801388 > [02:01] thanks. [02:02] i think perhaps the quickest fix is to have those links off by default and only enabled when needed. initially just for bugtask assignee [02:02] we can add other cases as and when required [02:02] wallyworld_: But jcsackett only made it global three weeks ago, as requested. [02:03] hmmm. but there's been some collateral damage :-) [02:04] certainly "Assign Me" only makes sense in certain contexts [02:04] "Choose Me" is better in other contexts [02:05] this is polish, and very visible polish at that [02:05] Indeed [02:05] wgrant: rabbit test ? [02:06] so, i think: default wording should be "Choose Me". bugtask assignee picker can override wording. select team member picker can disable links [02:06] but the links will remain on by default [02:07] and we can deal with other cases as the arise [02:07] they [02:19] lifeless: Doesn't help. I think something might be wrong with epmd. [02:19] rabbit says it starts fine, but the LP side of things can't see it. [02:23] wgrant: the outer rabbit ? [02:24] lifeless: By disabling the hardlink restrictions on the host the system rabbit now starts. But the LP rabbit does not. [02:26] Ah, and it's still broken without aufs. [02:28] wgrant: so I have similar symptoms [02:29] the host rabbit I have running and responding to rabbitmqctl [02:31] Status of node 'tmpMn00ML@lucid-test-lp' ... [02:31] Error: unable to connect to node 'tmpMn00ML@lucid-test-lp': nodedown [02:31] is what the test suite sees [02:31] Yes. [02:32] Status of node 'tmpp9GAiQ@lucid-lp-temp-uxeO' ... [02:32] Error: unable to connect to node 'tmpp9GAiQ@lucid-lp-temp-uxeO': nodedown [02:32] diagnostics: [02:32] - nodes and their ports on lucid-lp-temp-uxeO: [{rabbit,57084}, {tmpp9GAiQ,41334}, {rabbitmqctl1385,39815}] [02:38] So, only two aufs-specific problems now. And one of them is fixed by disabling the overzealous hardlink protection. [02:38] The other is that containers do strange things on reboot unless you remount. [02:39] Possibly they unmount too aggressively. [02:39] I had my pg corrupt itself [02:39] there is a known bug in upstart in lucid [02:39] I guess I should test the overhead of not using aufs. [02:39] See how it affects the cache. [02:40] Are LVM and the block cache friendly enough that snapshots will share cache? [02:40] AIUI - [02:40] new blocks written will be distinct [02:40] reads will be common [02:41] Let's use btrfs and see what happens. [02:42] I've found btrfs a little slow [02:42] anyhow [02:42] gl [02:57] Intriguing. [02:57] Same reboot issue with btrfs. [02:58] AIUI its a race condition, not fs [02:58] kill -9 too early [03:53] lifeless: Ah. [03:53] lifeless: Shutting down the guest remounts the root FS ro. [03:53] lifeless: Normally that fails, since there are files open rw on the host. [03:53] But with aufs/btrfs, the host doesn't usually have rw handles, so the remount succeeds. [03:54] wgrant: the guest fs ? [03:54] lifeless: Yes. But that is a tree in a host mount, so it takes the whole mount ro, it seems. [03:55] thats special; I haven't seen that happen to me [03:55] Are you running it on an FS that isn't your root fs? [03:55] no [03:55] Then you're always going to have files open, so the ro remount will fail. [03:55] its mount table has nothing but proc etc + the home bind mount [03:56] bug filing time ;) [03:56] Possibly. [03:56] dbus won't start on a btrfs snapshot :( [03:56] big loss :P [03:56] Well, means no avahi. [03:57] Which means I have to go hunting in dhclient's status files for the IP address. [03:58] Ah, stale pid got caught in the snapshot. [03:59] Anyone mind if the dogfood appserver gets a solid kick? [04:00] Kill it! [04:00] Even better if you're QAing your stuff :) [04:00] BTW I finally put LC_ALL in my mawson .profile… what a relief! [04:00] Yes! [04:00] LC_ALL=C, that is [04:00] It is a bit spammy otherwise. [04:00] I didn't dare do it for the launchpad user, but… [04:00] It doesn't matter. [04:01] It should persist through the sudo. [04:01] Shouldn't it? [04:01] It does. Just such a nuisance for everyone to have to do this. [04:08] Bah. [04:10] Bah indeed. That branch I needed to Q/A hasn't hit db-devel yet. [04:11] Maybe I'm pulling the wrong df branch. [04:11] I'm going to restart the df appserver again. [04:12] Well, this bah was because I accidentally deleted the root subvolume and not just the snapshots. [04:13] You accidentally the root volume? [04:13] Indeed. [04:13] 8 out of 10 doctors recommend not deleting your root volume. === wallyworld_ changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: wallyworld* (jtv) | Critical bugs: 211 - 0:[######=_]:256 [04:15] Hmm, Sunday is going to be warmish in Dublin, and the rest of the week is only slightly warmer than it is here. What summer! [04:16] hi wallyworld_! I won't be around much today because of travel stuff. [04:16] wgrant: and "here" means winter, right! [04:16] jtv: Right. [04:16] jtv: hi, me too. i've been in and out this morning getting stuff sorted [04:17] At the Wellington sprint I kept saying "winter" not because the hemisphere had me confused, but because the weather was so bad. [04:17] It was cold everywhere I went that winter, pretty much all around the world. [04:17] It went onto the books as the hottest winter ever. [04:18] the weather in wellington that week was really very horrible for the time of year [04:18] heck, it was pretty horrible for any time of year [04:19] Excuses, excuses. [04:19] Still. [04:19] Was better than the 46 we had back home during that week. [04:24] lifeless: Hmm, it works. [04:24] econtext [04:27] wgrant: ^ [04:28] Bah. [04:28] Thought testr run --parallel was working. [04:28] Just blew up. [04:31] wgrant: well, progress [04:31] we need to fix the mangling of stdout [04:34] This is getting far too annoying. Rebooting. [05:01] -> shoopermarket [05:02] "Shupper Shulloper" [05:02] (Not sure how prevalent 12th Man is) [05:17] Project db-devel build #662: STILL FAILING in 5 hr 33 min: https://lpci.wedontsleep.org/job/db-devel/662/ [05:19] Well there's a reboot I came to regret. Looks like today's update renders my MacBook unbootable! === michaelh1 is now known as michaelh1|away [06:01] lifeless: have you tried connecting to postgres using an sql client from outside the lxc container? [06:03] wallyworld_: O hai. https://code.launchpad.net/~stevenk/launchpad/dsp-vocab/+merge/65762 [06:03] * wallyworld_ looks [06:27] wallyworld_: I have not; you'll need to reconfgure postgresql to permit that [06:28] lifeless: any ftfm page you can point me to? if you don't have it handy, i'll search for it [06:28] rtfm [06:29] http://wiki.postgresql.org/wiki/Client_Authentication [06:43] Project devel build #833: STILL FAILING in 5 hr 42 min: https://lpci.wedontsleep.org/job/devel/833/ [06:48] How do we test that zope won't allow a user to access something? I'm trying assertRaises(ForbiddenAttribute, ) but that, er, fails with a ForbiddenAttribute exception. :-) [06:48] In doctests it's easy, but what about TestCases? [06:49] attrgetter? [06:50] yup, getattr does the trick. [07:05] wallyworld_: I think I'm calling SimpleTerm() correctly. And secondly, toTerm() isn't supposed to be called directly, in which case the query in searchForTerms would have warmed up the cache. [07:10] Yay, all containers 6 connected to testr... [07:44] lifeless: Much slower with btrfs, it seems :( [07:45] lifeless: But I've got testr integration mostly working now. [07:45] Seems to be running OK. [07:45] hopefully will take less than an hour :) [07:45] wallyworld_: Prod? [07:46] wgrant: \o/ \o/ \o/ [07:47] lifeless: (back to aufs, seems to be running quickly) [07:48] And a lot less memory pressure now it's not 6 instances of the one memory hungry test running at once. [07:56] Project parallel-test build #67: STILL FAILING in 1 hr 14 min: https://lpci.wedontsleep.org/job/parallel-test/67/ [08:04] hi rvba === jtv is now known as jtv-afk [08:29] Hmm. 12000 tests done, CPU idle... but it's just sitting there. [08:29] Upsetting. [08:32] StevenK: sorry, i was out buying dinner before. i'll take another look at it === jtv-afk is now known as jtv [09:13] wallyworld_: Did you have another look? I hate to be pushy, but I'd love to land this before my flight. [09:14] StevenK: started to but now i have to head off to soccer. i was going to finish when i got back but that doesn't help you. all the other cases of initialising the simple term i know of pass the object as the first param, hence my confusion [09:16] morning all [09:16] StevenK: i guess it depends on what you expect or want the vocab to represent [09:18] and how it's values will be used - that's the context i am missing [09:19] wallyworld_: So, since this isn't used any where, I don't think it matters much. It can be fixed when it gets hooked up the UI? [09:20] if you want. if i were doing it i would prefer to get the terms correct so that the next branch was cleaner. you need to get a +1 from jtv anyway, so maybe he can provide a view :-) [09:20] Was I invoked? [09:21] hi bigjools [09:21] jtv: i need to head off to soccer. there's a review that needs +1. StevenKwants to land it before his flight. i had some questions as to whether the SimpleTerm was being correctly constrcuted [09:22] StevenK: url? [09:22] jtv: https://code.launchpad.net/~stevenk/launchpad/dsp-vocab/+merge/65762 [09:22] I'm on it. [09:22] jtv: thanks. i'll have a look when i get back to see what transpired [09:51] bigjools: I just remembered — differences in how overrides were set were what led me to the fixes I made. It's just possible that (a) my branches fix the queue-tool overrides problem as well, (b) trying the overrides on sync jobs from the script now fails more spectacularly because of the security changes, or (c) the script doesn't set overrides at all somehow. I forget the details. [09:52] jtv: well let's see! [09:52] I did write you a lovely addOverride method [09:58] for which I thank you [09:58] (not saying how, but I do) [09:59] StevenK, wallyworld_: to cut things short I apprved, with remarks. [10:12] jtv: For your reference, there is a self.assertProvides() [10:12] the great thing about standards... [10:12] GAAAAAAAAAAAAAAAAAH [10:13] All that effort to enter objectives and *one* *lousy* *bad* *page* *load* undoes it all for me. [10:13] Which happen all the time. [10:13] And the page is constructed such that even if you have it all ready for copy & paste it's still a chore and a risk. [10:24] my external monitor has a bug [10:39] ok, I wish I had discovered bzr qdiff a long time ago, it makes reviewing much easier [10:46] wgrant: do you want to double check rvb's multi-parent branch? I am happy for it to land. [10:52] bigjools: I'm not here, but doesn't that need to use the copier if the archive is distinct and not empty? [10:53] Now it uses the cloner whenever the series is empty. [10:53] what did I miss ... [10:54] wgrant: what do you mean by distinct? [10:57] ah never mind [11:04] bigjools: Intra-archive copies can't produce file conflicts, inter-archive copies can. Regardless of series content. [11:05] yeah [11:49] allenap: thanks again for fixing up the doctest id to not have the 'tests/../doc' bit in it [11:50] allenap: makes running tests so much easier and eliminates an annoying bit of friction [11:50] jml: Hehe, I'm glad :) [12:25] Project db-devel build #663: STILL FAILING in 5 hr 41 min: https://lpci.wedontsleep.org/job/db-devel/663/ [12:47] Yippie, build fixed! [12:47] Project devel build #834: FIXED in 5 hr 52 min: https://lpci.wedontsleep.org/job/devel/834/ === bac changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: bac | Critical bugs: 211 - 0:[######=_]:256 === matsubara-afk is now known as matsubara [14:03] Morning, all. [14:03] Morning, abentley [14:03] abentley, shall we do a standup? Just you and I today. [14:03] deryck: morning. [14:03] deryck: sure. [14:03] cool [14:04] morning deryck [14:06] deryck, huwshimi: fwiw, it looks like landscape use lazr.anim and the lazr-js loader. [14:10] wiki page updated accordingly [14:12] jml: From memory we use lazr.anim but don't use the loader [14:13] jml, yeah, I think they use the green flash and spinner stuff. [14:17] deryck: ok. cool. I raised this because the wiki said, "they don't" and lifeless said, "they do", and I'd hate for us to get bogged down in a discussion about a verifiable fact. [14:18] jml, right. understood. [14:18] what is this wiki page we speak of? [14:20] deryck: The agenda for the Thunderdome [14:21] ah [14:21] deryck: O hai -- I have silenced Windmill on Jenkins. [14:21] StevenK, thank you sir! [14:26] jml, huwshimi -- looked at the page. I'm very confident that no one else cares if we abandon lazr-js. No one is supporting it, even if they use bits. [14:26] Burn! [14:26] Burn it. [14:27] U1 has abandoned it entirely, yeah? [14:27] wgrant: I believe so [14:27] wgrant: I just spent 10 mins finding out for landscape. I think it's someone else's turn to verify facts. [14:28] deryck: You have spoken to rockstar about lazr-js. Are they using it? [14:29] The only lazr-js references in the ubuntuone-servers tree are in test/lint configs. [14:31] huwshimi, we are not. [14:32] rockstar: OK, just checking. Do you have plans to use any of it in the future? [14:32] huwshimi, no. The first thing I did when rotating to U1 is to get rid of it. [14:33] rockstar: OK thanks :) [14:33] huwshimi, after I finish this (very large) project, I'll probably start focusing on bringing phazr in. [14:33] rockstar: Ah right. I thought you might have been using that already [14:34] huwshimi, well, in a way, we are. We're just not using it as a dependency. [14:38] but to be clear, the goals for phazr are much different... generic, extensible widgets, that each project can build in.... [14:38] the goal is for each project to do their own js development in their own tree, though. [14:38] rockstar, right ^^ ? [14:38] deryck, indeed. [14:39] huwshimi, to be clear, much of what is in phazr now is stuff that we have in U1, but are pushing to a place where LP et al. can use it. [14:45] rockstar: Right, got it. [14:45] Project parallel-test build #68: STILL FAILING in 1 hr 17 min: https://lpci.wedontsleep.org/job/parallel-test/68/ [14:50] huwshimi, ideally, I'd like to be able to push phazr stuff upstream into the yui3-gallery or yui3 as well. [14:51] rockstar: That sounds great. yui could do with some more contributors [14:53] hi cr3, your branch passed all of the ec2 tests last night but got kicked out of PQM due to testfix mode. it is in PQM right now. [14:54] bac: thanks for the update, when should it finally land in production? [14:57] rockstar, huwshimi -- I agree... we really should be thinking about yui as the dependency more. [14:57] and less about lazr or phazr. [14:57] cr3: it'll go onto qastaging.launchpad.net later today. when that happens the bug will be tagged with 'qa-needstesting'. at that point, you should verify that everything works as expected and the tag changed to qa-ok or qa-bad. [14:58] cr3: er, s/the tag changed/change the tag/ [14:58] deryck, indeed. lifeless' concept of owning the whole stack is important in this case. [15:00] hey [15:00] what's bugtask.priority for? [15:01] 2005 [15:01] lol [15:01] severity was renamed to importance, priority was apparently never deleted. [15:01] heh [15:01] what wgrant said ;) [15:01] Maybe it was 2006. [15:01] I'm no DBA, but wouldn't deleting it mean faster queries? [15:01] a while ago, anyway [15:02] also, there are bounty tables in the db [15:03] this stuff should be removed, yes. [15:03] I can imaging narrowing a huge table by a single int not necessarily being worth the downtime [15:04] *imagine [15:05] Well, it depends on the table :) [15:05] But BugTask.priority isn't BranchRevision.id. [15:05] bac: I should do that? will I get a notification or will the notification appear in the bug report? [15:06] cr3: you'll get bug mail when the tag is added [15:06] bac: awesome, thanks for the heads up, I get so much bug mail it's possible that I could miss it but I'll pay particular attention to that one [15:06] o/ [15:07] mhall119 and I need some help with LP API. [15:07] Or rather the optimal way to use the LP API. [15:07] shoot. [15:07] loco.ubuntu.com needs to know the members of all the teams under loco-teams [15:07] we need to get user data from everybody who's a member of https://launchpad.net/~locoteams [15:08] most will be indirect members [15:08] Does doing a +participants call make sense? [15:09] and does it overload LP too much. [15:09] especially for a team with 18370 members. [15:10] oh, and we'll need to keep our data updated, so we'll need to do this regularly [15:10] huwshimi: now officially at https://dev.launchpad.net/Projects/Rebranding [15:10] mhall119, nigelb: good question [15:10] * jml flees to a meeting [15:10] jml: Thanks [15:11] lol [15:11] haha [15:11] he's already fleeing from LP team ;) [15:12] jcsackett, do you have time to mumble for the minutes my computer is operating? [15:12] sinzui: sure. [15:34] my script tells me that we've closed 500 critical bugs since the start of the year and had 150 opened. that can't be right, can it? [15:37] Oh, its the friday before the sprint. jml will not get a pie on his face :( [15:38] ahh, I see. I'm not counting bugs that have been raised to Critical that were already opened. [15:40] hmm. [15:40] I guess I'd need to look at BugActivity to get that info [15:44] mhall119, nigelb: you may want to investigate using the team.membership_details method [15:45] mhall119, nigelb: it'll give you back a list that only includes a link to each member, not a fully populated person object [15:45] but, sadly, it doesn't account for subteams. you'd have to dive into each separately [15:46] ouch [15:46] doesn't team.participants do that diving for me automatically? [15:46] but you can get a list of those with team.sub_teams [15:46] nigelb: it does. but it returns fully populated person objects...each of which is pretty big [15:46] in our use case there is a team with only sub teams. [15:46] bac: that's going to be a lot of hits to LP [15:47] nigelb: if your structure is that flat, one team with one layer of subteams underneath, it would be pretty easy [15:47] so its either one big request to LP or a whole bunch of small requests (18000). [15:47] nigelb: no, it would be one request per subteam [15:48] ah, get a participants call in each subteam? [15:48] and the sub-team request would give us a list of populated person objects? [15:48] it's not going ot be quite that flat [15:48] as lots of loco teams have subteams [15:48] mhall119: no, it would give you a populated team. you'd then ask for membership_details on each of those teams [15:49] mhall119: and you might get a lot of duplicate persons who are members of multiple locos [15:49] I should probably try doing this with launchpadlib and see how it goes [15:49] bac: which is fine, we can ignore them [15:49] nigelb: if you do this: 'import httplib2; httplib2.debuglevel=1' before instantiating your lplib object, you can see the traffic being generated [15:50] oh, w00t, that would help us understand the calls better :) [15:53] mhall119, nigelb: stating the obvious, you could then keep a minimal reference to each person, find the changes, and only have to ask lplib for the full updates for the new people or those that leave. first time will be rough, though. [15:54] bac: the other thing we have to figure out make sure we don't count people who changed LP names and include old/new as different people. [15:54] nigelb: urgh [15:54] nigelb: we're getting the new django-openid-auth package ready that'll let us follow renames [15:55] mhall119: not in the gap when we do this. [15:55] worst case we'll have both old and new usernames in LD for a short time [15:55] yeah. [15:55] deryck, abentley I believe you can now get html5-browser 0.0.6-0 which works with 64bit natty [15:55] sinzui, yay! Will try here shortly. [16:02] sinzui: I get a bunch of Permission Denied: http://pastebin.ubuntu.com/631839/ [16:02] oh dear [16:02] abentley, I know that is the out-of-proc calling the pyc [16:03] I need to see if I lost code, or setup.py did something [16:04] abentley, was this on the first of second run? [16:04] sinzui: second run, I believe. [16:05] That is a little more comforting [16:05] * sinzui reads code [16:05] sinzui: the first run also failed. [16:05] oh, okay then [16:10]