[09:06] "ERROR failed to bootstrap model: cannot start bootstrap instance in availability zone "travis-job-28b2fcc6-d80d-4e62-8594-4e186717b5f0": not found" [09:06] this seems very strange [09:07] 2.7 branch... [09:09] there is really something weird in 2.7 branch, it has random lock ups in travis when attempting to bootstrap, I can replicate it locally though [09:46] manadart, got a sec [09:46] stickupkid, achilleasa: OTP with Atos. Gimme a few. [09:47] manadart, sure, take your time, just want to pick your brains [09:47] achilleasa, CR this one https://github.com/juju/python-libjuju/pull/342 ? [09:47] stickupkid: looking [09:48] I actually started going through that late yest but hit EOD... [09:56] stickupkid: HO? [10:29] manadart, i can't replicate locally, but i have seen it happen :| [12:50] manadart, should we be hitting this "juju.core.cache programming error, unit removed before being added, application name not found" [12:51] stickupkid: I am getting that consistently in the logs when I run 'juju remove-application' [12:52] stickupkid: No. That's an issue with the LXD profile watcher. [12:52] manadart, we should fix [12:52] stickupkid: Yes. [12:53] manadart, i fix [12:57] it's not really a programming error if we hit it all the time :| [13:59] morning [14:04] manadart, https://github.com/juju/juju/pull/10541 [14:04] stickupkid: Looking. [14:04] manadart, ta [14:24] manadart, i agree, i've clean up the method call [14:24] stickupkid: Cool. [14:28] stickupkid: I got the AZ error bootstrapping localhost. [14:29] manadart, https://github.com/juju/juju/pull/10424 [14:34] stickupkid: With develop, that didn't fix it for me. [14:34] manadart, something else has changed then :( [14:37] stickupkid: I think I see it. [14:37] manadart, nice, if i could replicate it, I'd have a nice fix up :( [14:37] stickupkid: Looks like the issue achilleasa was having too. [14:40] stickupkid: Looks like a change in LXD. Getting the image alias returns a not found instead of just a nil alias now. [14:40] manadart, ah really, but we've not bumped the client, that's unfortunate [14:40] stickupkid: I will check 2.6 too. [15:00] stickupkid: LOL; ineffassign fixes busted it. [15:00] manadart, haha where? [15:00] ho? [15:00] stickupkid: Sure. [15:21] stickupkid: https://github.com/juju/juju/pull/10542 [15:33] hml: recovering ok? [15:37] rick_h: coffee is my friend ;_D [15:37] :-D [15:37] manadart, just doing a final Q&A on the PR [15:38] rick_h, part 1 of updating deployment https://github.com/juju/os/pull/9 [15:39] stickupkid: cool ty [15:42] manadart, done [15:42] stickupkid: Ack. [16:39] is there any reason why model names can't have underscores, seems weird [16:52] stickupkid: got a few min to help me with a bit of lxd black magic? [16:52] achilleasa, sure [16:52] i like black magic [16:53] being juju and all [18:47] cory_fu: Found a breaking bug in layer:basic with `use_venv: False`: https://github.com/juju-solutions/layer-basic/issues/146 [18:47] tl;dr, any charm not using venv is broken [18:54] aisrael: Hrm. That's been like that for 4 years. Why has it only come up now, I wonder? [18:55] cory_fu: I wonder if it has to do with the recent changes around setuptools? I guess I could pull the previous revision and test against that. [18:56] or something more recent, I mean. Maybe something in setuptools itself changed. [18:56] aisrael: Well, it may just be the newer pip [18:56] Seems like it would be safe enough to just add --ignore-installed to that line? [18:57] cory_fu: I think that's a reasonable change [19:01] aisrael: I'm working on some OpenStack integration testing ATM. Would you mind throwing up a PR? [19:01] cory_fu: sure, not a problem [19:02] Thanks [19:54] aisrael: flake8 is complaining about your line length. :p [19:55] cory_fu: just noticed that :/ [20:30] cory_fu: lint fixed :) [20:31] aisrael: Thanks, merged