[09:09] gmb: morning, ready to pair when you are [09:09] Hi frankban. I'll be ready in about 20 minutes time; just need to send off some emails first. [09:09] gmb: cool [09:34] frankban, Okay, I'm ready. I'm in https://plus.google.com/hangouts/_/extras/canonical.com/subunittags (without video, but I'm sure you'll survive :)) [09:37] gmb: joining [11:05] 2-factor authentication is pissing me off [12:12] bac benji frankban gmb, apologies. Was trying to get the convoy packaging issue resolved, and now I'm blocked on a family issue. Hopefully I'll be ready in 3 [12:13] ok [12:15] bac benji frankban gmb call in 1 or asap [12:35] Rats. Should probably have stayed in the hangout... [12:35] face [12:36] gmb, you come up with such amusing hangout names, so I nominate you to come up with our zope:layer hangout [12:36] bac: you should get some tea and then we'll hang out [12:36] benji: :) [12:36] you know i was just about to type that [12:36] no pressure on the being amusing bit :-P [12:41] my mind reading must be working well today [12:41] gary_poster, frankban: https://plus.google.com/hangouts/_/extras/canonical.com/layeruponlayerofconfusion [12:41] very nice, though I am all alone there [12:46] bac: when the tea is brewed, the horde awaits [12:46] benji: yes, i had to make a new pot. i'm ready now. [12:47] benji: normal bat channel? [12:47] bac: yep [13:42] benji I meant to tell you on call: get MyUnity from software center and you can have a nice way to control size of workspaces, along with other stuff [13:43] gary_poster: oh yeah! I forgot about that. I actually have it installed but forgot that it existed. [13:43] heh cool [13:44] bac & benji, hey I'm sorry to interrupt but could you join us in https://plus.google.com/hangouts/_/extras/canonical.com/layeruponlayerofconfusion ? We have a solution, but the only explanations as to why it works don't make sense, so we'd like to see if either of you have insight. We can constraint to 10 minutes: over at 10:55 [13:45] constrain [13:52] bac: I think I'm here but I don't see you. [15:14] bac benji frankban gmb, fwiw I wrote some annoying instructions on how to get the list of workers from the subunit links on the ResultsLog page. The instructions are at the bottom of the page (https://dev.launchpad.net/ParallelTests/ResultsLog). Improvements & questions welcome. [15:16] gary_poster: benji will save us from all that soon i hope [15:16] bac, via the buildbot changes? Yes, though we may still need/want to be able to analyze the old subunit outputs [15:24] gary_poster: since you already have some context, you may be interested in the lp-lxc-ip MP: https://code.launchpad.net/~frankban/lpsetup/lp-lxc-ip/+merge/103502 [15:25] cool frankban [15:26] will look [15:26] thanks gary_poster [15:52] frankban, that branch is awesome! you have to talk me through some of it so I'm sure I understand what is going on. :-P https://talkgadget.google.com/hangouts/_/extras/canonical.com/goldenhordereview ? [15:52] gary_poster: joining [16:15] frankban, Is lpsetup now considered a useable replacement for rocketfuel-get, or should I consider to tell people to use rf-g for now? [16:39] gary_poster, Do you know the answer to that ^^? [16:45] gmb: lpsetup is not ready to replace rocketfuel [16:45] frankban, Okay, cool. Thanks. [16:59] frankban, I am running sudo nosetests. (at least) one test takes a really long time, right? I should not assume that it is hanging? [17:00] gary_poster: yes, it creates an lxc, and the first time it could take some minutes [17:02] ok cool [17:04] frankban, fwiw http://pastebin.ubuntu.com/945997/ [17:04] trying again [17:05] gary_poster: hum, do you have the lxcbr0 interface up? [17:05] yes frankban [17:05] and failure repeated [17:07] gary_poster: could you please print the ip before the failing lines and the run nosetests -vs? [17:08] sure [17:10] frankban, it passed now :-/ [17:10] 10.0.3.6 fwiw [17:11] ah interesting frankban [17:11] frankban, I ran it again without -vs for the heck of it [17:11] I got one failure (test_get_ip_ [17:11] ) [17:11] and the ip was 10.0.3.182 [17:12] gary_poster: this is another test run, right? [17:12] I did it again and it passed [17:12] then I did it again and test_restart failed [17:12] with ip 10.0.3.28 [17:12] yes frankban,separate test runs [17:14] gary_poster: the ip is retrieved correctly, maybe this is a race condition, could you please try to add time.sleep(1) before the failing lines? [17:15] gary_poster: maybe the dns needs a second to register that ip, and maybe I should just test that the ip is actually an ip, rather then using socket.gethostbyaddr [17:16] ok [17:18] frankban, with time.sleep(1) I got both failures :-P [17:18] 10.0.3.38 in both cases [17:19] I'll increase sleep for the heck of it [17:19] 1 full pass with 5 seconds [17:21] another one... [17:22] frankban, I entered my review on the MP. It has a new thing or two, so take a glance if you would. [17:22] another full pass... [17:22] another one... [17:23] another... [17:23] thanks for the review gary_poster. so, it seems a race condition [17:23] frankban, time.sleep(5) seems to do it, yeah. [17:23] gary_poster: our buildbot run was bit by the _buffered_calls AttributeError [17:23] (another one) [17:24] benji, really! super annoying [17:24] gary_poster: in that case I have two choices: [17:24] gary_poster: yeah, we're going to have to fix that before we can continue [17:25] ok benji. good luck. lemme know if you'd like another pair of eyes. I cc'd you on the email to jml about it, yeah? [17:25] no reply [17:25] 1. generalize the retry decorator and use it with socket.gethostbyaddr [17:25] yep [17:25] 2. just test the the returned value is a valid ip [17:25] 3. (potentially painful) figure out why there is a race condition and what you can wait for instead [17:26] I kind of prefer 1 to 2 frankban, but will defer [17:27] gary_poster: i prefer 1 too, for 3, I have to reproduce the race condition (never seen before) in my machine [17:29] frankban, sure. go for #1 [17:31] hi benji [17:34] bac: take a look at this http://ec2-23-20-226-202.compute-1.amazonaws.com:8010/builders/lucid_lp/builds/0/steps/shell_9/logs/stdio [17:34] * bac looks [17:36] benji: you've encountered this problem before? [17:37] yep, we reported to jml but he hasn't had time to look at it; I guess it's our turn now [17:37] I'm filing a real bug for it now. [17:41] gary_poster: I'be just pushed a revision that uses retry for those tests [18:04] frankban, passed fully three times in a row. +1 [18:05] benji: i'm doing a review...ping me when ready [18:06] bac: ok [18:18] bac: https://talkgadget.google.com/hangouts/_/extras/canonical.com/goldenhorde when you're ready [18:18] benji: ok, give me a bit [18:47] benji: joining [18:47] k [19:52] benji, what is the bug number of the buffer attribute problem you filed for testrepository? (I'm assuming you are working on it, and will update the board as such) [19:55] bug 988481 [19:55] <_mup_> Bug #988481: Testr tests generate an AttributeError when run with latest testtools < https://launchpad.net/bugs/988481 > [19:55] gary_poster: yep, that's it [19:55] thanks [19:55] it's melting our brains at the moment, but when life gives you melted brains, you make brain soup [19:59] heh [21:05] benji and bac are listed as Ubuntu contributors for 12.04! [21:05] heh [21:05] So I hear from sinzui [21:05] for why? [21:06] lxc? [21:06] lxc maybe? [21:06] juju? [21:06] maybe [21:07] gary_poster: I'm listening to the song listed under "Latest Single" on the lower right of http://perhapscontraption.co.uk/ and it made me think of a group that you like that I can't remember the name of :) [21:07] :-) on call, will check after [21:08] benji: type 'yes' [21:08] heh [21:08] thanks bac [21:08] np [21:08] i felt so helpless [21:09] lol