[00:02] lifeless: So is 'irony' just a metal to you? :-P [00:09] I think the boron sell it? [00:29] lifeless: Hm [00:30] I think we can speed up the test suite significantly by completely disabling the lazr.restful representation cache. [00:30] We don't read from it, but we still invalidate it on every object flush. [00:30] Saves 25% on makePerson. [00:31] bug 780835 [00:31] <_mup_> Bug #780835: representation cache a pessimization < https://launchpad.net/bugs/780835 > [00:32] Sure, but it doesn't say that even when disabled (which is everywhere) it still hurts writes. [00:32] wgrant: feel free to add that (or just rip the cache out) [00:33] I'm just going to rip it out, since reading from config there is probably too slow. [00:33] lazr.config is pretty awful. [00:34] (seriously, just profiled) [00:37] That saves 40% now, both in and outside the profiler. [00:40] o/ [00:40] Our Storm tracers are slowish too, but not so bad [00:40] they need to be replaced [00:40] there is a cleaner version in storm trunk [00:40] I context switched at the wrong point [00:41] Ah [00:41] Timeout checking goes through lazr.config [00:42] Oops [00:42] only until its overridden by the feature flag IIRC? [00:42] * StevenK goes next door to relocate a lawn mower ... [00:43] lifeless: Ah, true, so it only happens in the test suite I guess. [00:44] wallyworld__: How is the SSD? [00:45] StevenK: not sure yet. i downloaded a precise iso and am yet to install it [00:45] wgrant: check the implementation if you like [00:45] gotta decide if swap should live on the sdd r not. opinion? [00:45] wallyworld__: swap? what swap. [00:46] s/ should live on the sdd r not// [00:46] lifeless: guess you have lots of ram? [00:46] wallyworld__: 8GB on my laptop, 16GB on my desktop. [00:46] i have 4BB [00:46] GB [00:46] wallyworld__: should be able to do quite a lot w/out swap [00:46] Oh, right, and you run LP on amd64. [00:46] I have swap on the SSD, but that was so I could hibernate. Which doesn't actually work anyway, so I could kill it. [00:47] yeah, i do run lp on amd64. it's not too bad [00:47] fsvo too [00:47] I have 1GB swap just-in-case [00:47] Swap: 1050616k total, 0k used, [00:47] i'm afriaid if a put it on the ssd it will kill the ssd [00:48] well, shorted it's life [00:48] significantly [00:48] only if you use it [00:48] what SSD did you get? [00:48] * StevenK peers at jenkins [00:49] crucial M4 [00:49] $180 on amazon [00:49] 128GB [00:49] yeah, and i will be using swap, so best to leave on the hdd i guess [00:50] if you're using swap, andhave an hdd available, definitely put it on the hdd [00:50] also [00:50] http://www.tomshardware.com/news/Crucial-m4-Firmware-BSOD,14544.html [00:50] * wallyworld__ looks [00:51] lifeless: yeah, saw that issue. i think the drives shipping recently already have the updated f/w but need to check that. [00:57] wallyworld__: anyhow, I don't know the crucial arch; the intel one its fine to run swap on - they have (IIRC) 30% slack space to wear level across [00:58] ok. thanks for the info. i'll do some research :-) [00:59] but that said, IMO when you hit swap you've already lost [01:00] 2.6 and 3.x are swappyish kernels anyway. They will try *really* hard to swap out at least a little bit if they can. [01:00] StevenK: I beg to differ :) [01:00] uptime [01:00] 14:00:30 up 9 days, 18:34, 6 users, load average: 0.05, 0.12, 0.10 [01:00] uname -a [01:00] Linux lifeless-64 3.2.0-8-generic #15-Ubuntu SMP Wed Jan 11 13:57:44 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux [01:00] lifeless: Where a little bit is like a few pages [01:01] Swap: 1050616k total, 0k used [01:01] StevenK: yes, I got that; but I don't /see/ it :) [01:01] steven@liquified:/media/ECD5-838E/tv% free | tail -n 1 [01:01] Swap: 3903484 5184 3898300 [01:02] free | tail -n 1 [01:02] Swap: 1050616 0 1050616 [01:02] wow, a geek dick measuring contest [01:02] someone said dick? [01:02] "my swap is smaller than yours" [01:02] anyhow, a few pages won't break an ssd under any circumstances [01:02] No, I think lifeless just doesn't believe me [01:02] thumper: yes, of course it was me [01:02] StevenK: I accept you have a few pages swapped out [01:02] hi wallyworld__ [01:03] g'day thumper. you make it home ok? [01:03] StevenK: its had 9 days to find such a page for my laptop, and hasn't. [01:03] we had a "small" delay [01:03] wallyworld__: eventually [01:03] you delayed too? [01:03] how was the opera? [01:05] wgrant: that reminds me, i got a $300 Qantas voucher because of the delay. did you get one? [01:05] I did [01:05] StevenK: oh yeah, forgot you were delayed [01:05] NFI what the heck I'm going to do with it [01:05] come to brisbane :-) [01:06] Perhaps my wife and I will visit her family in Adelaide again [01:06] StevenK: adelaide?! [01:06] it says it's not transferable :-( [01:06] so i don't think you can split it up or share it with anyone else [01:07] wallyworld__: I was planning on buying 2 return seats. They shouldn't have a problem with that. I hope. [01:07] hope not. if you do, let me know how it goes [01:07] lifeless: And? [01:08] StevenK: I thought they were sydney-folk [01:09] lifeless: Sarah and her family are all from Adelaide. [01:09] wallyworld__: Bloody hell. $275 to fly to ADL on Qantas. One way. [01:09] wow [01:10] bit of a rip off [01:10] A bit, you say? [01:10] That's like calling a fish 'a bit wet'. [01:11] Or $380, if I choose to go via Melbourne. [01:11] May as well send me via Perth [01:11] StevenK: interesting! I wonder why I thought she was from Sydney [01:12] wallyworld__: the opera as a little surprising [01:12] wallyworld__: I was jet-lagged, and the opera theatre was warm and dark [01:12] wallyworld__: the normal naps and snaps occurred [01:12] oh no, you fell asleep [01:12] wallyworld__: I feel sorry for the people behind us [01:12] hah [01:13] wallyworld__: not fully asleep, just nodded off a few times [01:13] wallyworld__: I wasn't the only one though [01:13] it's the head snaps that give it away :-) [01:13] I can recall doing that in dark lecture halls when I was at uni ... [01:13] wallyworld__: I found the opera was better after half time when I read the storyline someone had printed out from wikipeadia and stopped trying to read the subtitles [01:13] lol [01:14] and how were the meetings? everything sorted out? [01:16] wallyworld__: I did too [01:18] too bad $300 doesn't get very far on Qantas [01:19] it gets you on the plane [01:19] Sometimes. [01:24] Haha [01:26] Sigh. Can we kill polls yet? [01:27] Now I wonder what could have triggered that? :) [01:27] I wonder. :-P [01:31] what happened with polls? [03:13] wgrant: I have to admit, I'm a little worried adding combobuild into build. [03:15] StevenK: Check the qastaging, staging and production build scripts. If they don't use build, you're safe enough. [03:16] (enough) [03:16] wgrant: Ah, but where the scripts? [03:20] StevenK: staging_restore.sh is in lp:lp-staging-scripts [03:21] qastaging-update.sh is elusive [03:21] And an old version of the production deploymgr config is in carob:~wgrant/dpm-lp-configs [03:22] * StevenK stabs SSO, and then twists the knife. [03:22] connection resets? [03:23] ssh launchpad@asuka "make -C /srv/staging.launchpad.net/staging/launchpad-new build LPCONFIG=staging" >> $LOGFILE 2>&1 [03:23] No build for me [03:23] Ah, but that's asuka. [03:24] Exactly where you want it. [03:24] It does it on gandwana too [03:24] That's probably because it was copied from asuka; not because it's necessary. [03:24] build: compile apidoc jsbuild css_combine sprite_image [03:24] The only thing gandwana needs from that is compile. [03:26] Right, so I want to branch this and set a CONVOY_ROOT [03:26] Probably [03:26] You also need to obtain qastaging-update.sh [03:26] And possibly a more recent deploymgr config from prasé c/o spm [03:30] Mornin' [03:49] wallyworld_: You don't show up in mumble as 'harshal-sengar' any more, right? [03:50] StevenK: not that i'm aware of [03:50] wallyworld_: You have an open RT (#43809) about that issue, you could ask IS to close it. [03:50] <_mup_> Bug #43809: missing manpage for wcsftime < https://launchpad.net/bugs/43809 > [03:50] ok [05:13] wallyworld_: Can't you just use an 'LP' namespace or something? [05:14] Or is that too global for deryck/rick_h's liking? [05:14] tried something like that, doesn't work across instances [05:14] Just for the event. [05:14] :( [05:14] atm, i'm devoid of ideas [05:18] * wallyworld_ goes to get the kid from school and transfer rego on "new" car. old faithful blue one finally died :-( [05:50] :( [05:50] Removing the representationcache may only have saved 5-10 minutes :( [05:50] It's still a plus [05:59] wallyworld_: Is bug #923973 the event issue? [05:59] <_mup_> Bug #923973: sync page is broken < https://launchpad.net/bugs/923973 > [05:59] i didn't raise that one, let me look [05:59] I assume the AJAX log and MP longpoll breakage is also that. [05:59] But those two are less important, as they only affect us. [06:00] wgrant: i'm not familiar with that bit of code. i raised bug 923619 to cover cases like editing a recipe, editing a blueprint, changing a branch name [06:00] <_mup_> Bug #923619: client model changed events broken < https://launchpad.net/bugs/923619 > [06:01] wallyworld_: I imagine they'll need separate fixes. [06:01] i don't know if the breakage in the bug you posted uses the cahce model events or not [06:01] or what meachanism is used, but sounds similar to what we're seeing elsewhere [06:02] It doesn't use those, no. [06:02] but if it does use Y.fire in some way, shape or form it most likely is now broken [06:03] just a guess [06:03] That was my assumption. [06:04] it depends on if separate yui() instances are used [06:05] but if there's a few breakages like this, we need to get this fixed asap [06:06] and two of the current maintenance squad were involved in the yui work at the epic, so they are well placed to look at this [08:56] good morning [09:42] #launchpad-kitchen === gmb changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: gmb | Firefighting: - | Critical bugtasks: 4*10^2 === matsubara-afk is now known as matsubara [11:42] AttributeError: 'TestPullerMasterIntegration' object has no attribute '_lock_actions' [11:42] I thought bzr 2.5 fixed that from happening? :-( [11:44] StevenK: I think it's mostly got to do with hook installation [11:52] I just edited a milestone inline on a bug page and the new impage where the edit icon should be was a 404 [11:52] with this URL: https://bugs.launchpad.net/testtools/+bug/881052/undefined [11:52] <_mup_> Bug #881052: assertRaises gives error about a lambda < https://launchpad.net/bugs/881052 > [12:30] bah, missed him === matsubara is now known as matsubara-lunch === Ursinha` is now known as Ursinha_ === Ursinha_ is now known as Ursinha [14:41] abentley, adeuring, rick_h -- standup in 5 cool? [14:42] deryck: k [14:42] deryck: fine for me [14:42] deryck: sure [14:47] rick_h, adeuring, abentley -- having issues connecting to mumble for some reason. [14:47] deryck: same here. [14:48] rick_h, adeuring -- did you guys get in? We may need Skype for abentley and I. [14:48] yea, we're both in [14:48] deryck: yes [14:49] hmmm, yeah, I can't get in. Doesn't like my password. [14:50] I got in. [14:50] adeuring, rick_h, abentley ^^ [14:50] heh, adeuring and I jumped out. adeuring went looking for a machine with skype installed [14:50] and abentley just got in! [14:50] ok, back to mumble! === matsubara-lunch is now known as matsubara === _mup__ is now known as _mup_ === almaisan-away is now known as al-maisan [15:51] mrevell, are we supposed to have a call in 10 minutes? I didn't get any invites for it [15:52] my calendar just warned me but I didn't see any invites before this [15:53] Hey salgado, yeah I sent an invite. Is that okay with you or do you want to reschedule? [15:53] mrevell, it'd be nice to have mabac and danilos as well. let's see what danilos thinks [15:54] danilos, ^ [15:54] Okay, great. [15:55] mrevell, did you send the invite to them as well or just me? [15:55] salgado, only you [15:56] salgado, mrevell knows me all to well to want to talk to me [15:56] :) [15:56] I know, but I thought you'd like to talk to him ;) [15:58] danilos, so, if you don't want to join I think I'll do it now and fill mabac later [15:59] salgado, mrevell: I'd be happy to join in, mumble on canonical server? [15:59] Sure or Google Hangout [16:00] ok, I'll be back in 1 minute [16:00] mrevell, salgado: let's do mumble, I am not yet comfortable enough with hanging... out [16:01] mrevell, can you hear me? [16:02] I hear you [16:02] I'm having some trouble with the mic [16:02] yes [16:04] unmute me! [16:04] salgado, can't you unmute yourself? [16:18] mrevell, https://blueprints.launchpad.net/linaro-status-website/+spec/individual-engineer-upcoming-work-view-mockup [16:27] lifeless: do you have a tabular schema for subunit data anywhere? === salgado is now known as salgado-lunch === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away === dpm_ is now known as dpm === salgado-lunch is now known as salgado [17:47] jml: maybe? [17:48] lifeless: thinking of writing a subunit -> csv filter to make data analysis easier. [17:48] ah [17:48] lifeless: want to try to stop me? [17:48] not at all [17:48] :) [17:48] subunit is streaming after all; so either casting it to a static form, or using a streaming language -> either make a lot of sense [17:49] the lp results tracker may have some inspiration, as may the linaro thingy [17:49] thinking id,start_time,stop_time,status for now [17:49] lifeless: do you know where I can find those? [17:52] somewhere on LP? lp:launchpad-results is one [17:55] * bigjools waves at lifeless the insomniac [17:56] o/ === mtaylor_ is now known as mtaylor [20:15] wallyworld_, you around? === matsubara is now known as matsubara-afk [20:25] * deryck needs to get away from the screen, back soon === deryck is now known as deryck[lunch] [20:54] deryck[lunch]: hi, ping me when you get back from lunch. i have to drop the kid at school soon, won't be long === deryck[lunch] is now known as deryck [21:08] wallyworld_, back when you are. === salgado is now known as salgado-afk [21:11] deryck: give me 25 minutes. i'm just leaving to drop the kid to school, be back as soon as i can [21:11] is there a way in launchpadlib for me to get a list of bugs that have a patch attached to them? [21:11] wallyworld_, np. [21:13] mhall119, use searchTasks and pass in the has_patch argument. see: https://launchpad.net/+apidoc/ [21:13] thanks deryck [21:13] np [21:43] deryck: hi, mumble? [21:45] wallyworld___, hey. give me 3 or 4 minutes to wrap what I'm doing and I'll meet you there. [21:45] ok [21:45] wallyworld___, and what's with all those underbars? ;) [21:45] no idea [21:45] NickServ says my nick is temporarily unavailable [21:49] wallyworld___, my brother in law came in sorry…. 5 more minutes [21:49] np [21:51] wgrant: i installed precise on a clean system. debversion for postgres 8.4 wouldn't install because libapt-pkg.4.10 wasn't there (precise uses 4.11 ). i can't find a deb for 4.10 anywhere. any hints? [21:53] wgrant: to get it working, i forced debversion to install and symlinked the required .so files to the 4.11 ones. but now apt insists on uninstalling debversion since it thinks it's broken [21:54] wallyworld___, ok, free now. firing up mumble [21:54] ok [21:54] wallyworld___: According to gary on the ML, LP on precise is broken in other ways too. I'll do a fresh LP installation today and sort it out. [21:54] wgrant: thanks [22:25] deryck: for the single instance name, a few people have expressed a dislike for "LPS". perhaps "LP_YUI" or something [22:29] wallyworld___, sure, I don't care about that. I'd like to avoid underbars though. [22:29] deryck: i have no opinion, just relaying someone else's :-) [22:29] wallyworld___, what's the dislike of LPS? Just a weird name? [22:29] yeah [22:29] yeah, I can see that. [22:30] np, leave it to you to decide :-) [22:30] ah, I don't care. I'll change it but drop the under bar. LPYUI work? Or LPUI? Or …? [22:30] :) [22:30] Launchpad Production Status [22:30] ^ LPS already exists, lets not have two. [22:31] lifeless: it used to be LPS, we are deciding on a new name [22:31] I think it was originally meant to be LPJS, but the sed script was wrong, and no one noticed or complained til now. ;) [22:31] thanks! :) [22:31] deryck: you serious???!!!! [22:31] wallyworld___, I am. ;) [22:31] lol [22:31] heh, I know. [22:32] LPYUI sounds ok to me, but meh [22:32] yeah, I'm not passionate about it either. [22:32] Just too lazy to type an underbar all the time. [22:32] * wallyworld___ nods [22:33] I think we should try again for LPJS and see if we get it right this time. ;) [22:33] LPJS sounds okay to me [22:33] Better than LPS anyway [22:34] I'm just kidding around. Any other name is fine with me. Will let rick_h pick since he'll be doing the sed-based branch tomorrow. [22:53] Later on, everyone. === spm` is now known as spm