[00:52] lifeless: Branch attached, please review. [02:01] jkakar: oh hai [02:01] jkakar: thanks [02:02] jkakar: if you're still around, theres another thing that might be useful :) [02:02] jkakar: would like you're input on https://bugs.edge.launchpad.net/launchpad-foundations/+bug/618019 [02:02] <_mup_> Bug #618019: OOPS may be underrepresenting storm/sql time [03:48] lifeless, I partly disagree, but you're welcome to explain why some other time. [03:49] * jml off to bed. [04:48] jml: bed! wow, little late :) [20:22] moin moin [21:05] hmm [21:06] where to file 'sprint' bugs ? [21:16] lifeless: the same place as blueprint bugs, though I can't remember where that was... [21:16] lifeless: 'blueprint' [21:17] thanks [21:17] I' [21:17] ll go back and find it in a minute [21:56] can LP API's be used in GAE ? [21:59] I can't think of an obvious reason why they shouldn't be, but I don't know. [22:02] jelmer: https://bugs.edge.launchpad.net/soyuz/+bugs?field.tag=timeout [22:02] :) [22:09] that's quite a few.. [22:10] I've been filing the next set - timeouts by % of request rather than volume per day [22:21] lifeless: I'm looking at that binary chop thingy [22:22] \o/ [22:22] it will be a bit fiddly I suspect :( [22:22] if I just run the LaunchpadZopelessLayer tests I get the failure [22:22] if its any consolation its trivial in bzr :) [22:22] thumper: \o/ ok thats a good starting point [22:22] so instead of starting with everything... [22:22] yeah [22:22] do the ls [22:22] so how do I run the test suite with the right output? [22:22] it will include the layer setup and teardowns [22:22] my default test run isn't subunit [22:22] so you can use that to shrink things [22:23] --subunit [22:23] cat .testr.conf - may give you some hints :) [22:24] running now [22:24] takes about 40 - 45 mintues [22:25] also I my first chop is going to take most of the tests out, as I think it is an interaction between the code import tests and the translation tests [22:25] * thumper goes to make first coffee [22:30] jelmer: wth does Distribution:+archivemirrors do that it ever times out :) [22:31] It might be the freshness queries. [22:31] In fact, that's almost definitely the case. [22:31] ... [22:31] It needs to aggregate freshness data for all series into a single value for each mirror. [22:32] in the webapp transaction ? [22:32] IIRC. [22:32] we only have what 10 series [22:32] * lots of archs and pockets [22:32] It's most likely the freshness stuff. [22:32] And I bet the queries aren't efficient... [22:32] ok [22:32] well, if you want a ++oops++ on it, shout === almaisan-away is now known as al-maisan [22:37] wgrant: distroseries:+builds is still pain [22:37] /msg nickserv identify OafNem1Glewg:Ter [22:37] lifeless: :( [22:37] al-maisan: time to change your password [22:38] yack [22:38] wgrant: nearly flat spread between 0 and 5 seconds [22:38] sorry [22:38] 0 and 7 seconds [22:38] Hm :/ [22:38] and then a 9% spike @ 15 seconds [22:38] lifeless: thanks for telling me anyway :P [22:38] al-maisan: might have been a failed script [22:38] al-maisan: wanted to be sure you knew :) [22:40] wgrant: https://bugs.edge.launchpad.net/soyuz/+bug/618405 [22:40] <_mup_> Bug #618405: DistroSeries:+builds timing out in ~9% of requests by edge timeout [22:58] jelmer: you might like lp:python-fixtures === al-maisan is now known as almaisan-away [23:09] I'd value a second opinion on bug 618019 [23:09] <_mup_> Bug #618019: OOPS may be underrepresenting storm/sql time [23:18] lifeless: looks interesting [23:19] lifeless, where do you find the time to think of, implement and release a new project each weekend ? :-) [23:19] morning [23:19] jelmer: that one has been percolating for ~ a year, I think. [23:19] hi mwhudson [23:20] 'morning mwhudson [23:21] lifeless, heh, ok [23:21] jelmer: I just hadn't written any code or anything [23:21] it simply needed actually getting a sunday with no tasks, no chores, no crises, no interrupts. [23:21] and voila [23:22] lifeless: did your recent experiences with launchpad layers have anything to do with it? [23:22] no [23:23] testresources has some api tensions [23:23] which I've talked with poolie and jml extensively about [23:23] in the 'wish it was better' sense :) [23:23] I concluded (dunno when ;P) that it was because it was doing too much [23:23] it should be a manager for an existing concept [23:24] rather than a whole concept + manager [23:24] so this project is targeted at that core concept [23:41] ah, ok [23:41] anyway, it looks interesting, I've put it on my list of stuff to look at :-) [23:41] :) === almaisan-away is now known as almaisan === almaisan is now known as al-maisan [23:48] hi there === al-maisan is now known as almaisan-away [23:54] 'morning poolie [23:56] * mwhudson finds http://www.kernel.org/pub/software/scm/git/docs/git-notes.html [23:56] jelmer: something for bzr-git roundtripping?