[00:12] Bug #1522948 opened: ensure-availability on rackspace spawns endless machines

[00:24] Bug #1522948 changed: ensure-availability on rackspace spawns endless machines

[00:30] Bug #1522948 opened: ensure-availability on rackspace spawns endless machines

[01:08] jam: http://juju-ci.vapour.ws/job/run-unit-tests-mongodb3/ was not super health when last run [01:18] axw_: I believe what nate found is a bug where DestroyVolumes is getting something other than what it expects in term of volume identifier :( [01:20] sinzui: quick +1 on my licence foo? https://github.com/juju/utils/pull/182 [01:32] sinzui: kvm-slave isn't resolving for me === urulama_ is now known as urulama [07:49] Bug #1523783 opened: Bootstrap failures: ERROR upgrade in progress - Juju functionality is limited

[07:52] Bug #1523783 changed: Bootstrap failures: ERROR upgrade in progress - Juju functionality is limited

[07:55] Bug #1523783 opened: Bootstrap failures: ERROR upgrade in progress - Juju functionality is limited

[09:06] frobware: morning [09:06] frobware: we going to do standup today? [09:06] frobware, morning! [09:07] voidspace, can do. I was trucking along with bridge per active NIC [09:07] cool [09:07] ditto on the apiserver side of the new facade [09:07] first things first though [09:07] coffee [09:07] voidspace, want to meet now? [09:07] ah [09:08] :-) [09:08] gimme 10 [09:08] I'll ping you when I have coffee [09:16] frobware: ok [09:17] voidspace, in the HO [09:17] omw [09:23] heyz and good morning, I thought you're in SFO/OAK right now? *stunning* [09:43] voidspace, any comments on https://github.com/frobware/juju/blob/master-multi-bridge-support/provider/maas/n.py would be appreciated. It's not the final name, just the bridge per NIC variant as we discussed in the standup. [09:43] frobware: looking [09:45] ty [09:45] it's *so* much nicer than the bash [09:46] I actually understand what it's doing... [09:46] Bug #1523837 opened: juju add-relation blocked Waiting for agent initialization to finish [09:47] the interesting stuff is in bridge_iface [09:48] frobware: it doesn't seem like it would be resilient against broken / strange eni formats [09:48] frobware: but I'm not sure that's possible, we're not trying to solve the general case [09:48] voidspace, agreed. [09:49] frobware: for example, the split on line 190 will break if there are not exactly four results from the split [09:49] voidspace, because the /e/n/i we parse is machine generated. [09:49] right [09:49] voidspace, I wasn't attempting a generic parser; for that you would really have to make it token based. [09:50] yeah [09:50] and it would be better to find one already built [09:50] but then bundling dependencies in cloud-init becomes problematic [09:50] so better to avoid going there if we can [09:51] frobware: it looks good to me [09:51] frobware: I can't see any problems with it (not there aren't any - just I can't see them!!) [09:51] frobware: and the Python is good, very readable [09:52] voidspace, the tests are still in Go which helps prove the generation of the script that gets run on the bootstrap host. [09:52] I would separate out the library and the script part so the library part can be unit tested [09:52] but maybe just black box testing the whole thing is sufficient [09:52] frobware: right, cool [09:52] and if we test the in/out of all the cases we know of then it's sufficiently tested [09:53] if it grows more we may want to unit test the components [09:53] voidspace, I was very conscious I wanted to test the interaction in Go, hence leaving them as Go unit tests. Ordinarily, I would have added python-based unit tests. [09:53] voidspace, for me there's two parts; re-rendering and making the interfaces active. [09:53] voidspace, the latter is pointless as a unit test IMO [09:54] it's still essentially a text transformation, so a black box test works well [09:54] right [09:54] voidspace, too many things on the target side can change; ubuntu series, kernel version, newer packages, et al. [09:54] needs to be a functional test [09:54] fun [09:55] voidspace, right. I think we need those functional tests. I just don't see any point mocking the fact that the pseudo interface became active. it will always pass or fail - depending on what you want. [09:57] frobware: agreed [09:57] voidspace, I think there are things we can do at the end of the script to verify that the default route still works, for example. [09:58] frobware: you could do that from Python too [09:59] voidspace, sure. the whole thing would be python based now. [09:59] there seems to be a TheMue filter :D [09:59] TheMue: o/ [09:59] voidspace: o/ [09:59] voidspace, it doesn't seem possible to rely the return value of ifup(1) or ifdown(1). [09:59] TheMue: have you started your new job yet? [09:59] frobware: ah... that's a nuisance [10:00] frobware: parse the results of ifconfig... [10:00] I'm only wondering that there's no sprint in OAK as it seems. [10:00] TheMue: the sprint is on, frobware and I aren't attending though [10:00] voidspace: Yep, sitting here creating a scalable message store in erlang. [10:00] TheMue: I have a nine month old baby and a tired wife and frobware has a wife going into hospital [10:00] TheMue: everyone else is there [10:00] TheMue: cool, you enjoying it? sounds like a challenge [10:01] voidspace: ah, ic. because I really would like to be there too. I love SFO. [10:01] TheMue: I'd love to be there too [10:01] voidspace: good reasons to stay at home, definitely. [10:02] yeah, unfortunately [10:02] voidspace: yeah, it's a challenge, especially the timeline for a first release :D technologically it's interesting and I reuse several experiences we've made during Juju development [10:02] cool [10:03] TheMue: a challenge is not a bad thing! glad your experience is proving useful [10:07] voidspace: definitely, I enjoy it. looking forward the numbers (of users, throughput, etc) when it is running in production [10:25] Bug #1517258 opened: juju 1.24.7 precise: container failed to start and was destroyed

=== benji is now known as Guest42305 [14:31] beisner, hiya! do you still plan to try with wily? [16:24] frobware: are you still looking @ bug 1523608? [16:24] Bug #1523608: Timed out waiting for device sys-subsystem-net-devices-juju-br0.device

[16:26] frobware, indeed. cycling T & V atm. interestingly, i'm now having T bootstrap success, still V fails. W is up next. [16:41] frobware, W bootstrap succeeded, V failed. but i get this trying to deploy a charm to the W enviro. is there a known issue?: [16:41] ERROR cannot upload charm: Post https://10.245.168.13:17070/environment/c08eab82-d996-4e0f-89df-97ed329a3b3e/charms?series=wily: EOF [16:43] frobware, ok interesting. i waited a bit, then re-attempted juju deploy, and it uploaded and started off to the races now. i'll not get distracted with that issue atm (needed to wait), in the interest of getting our lab back up. [17:17] Bug #1523783 changed: Bootstrap failures: ERROR upgrade in progress - Juju functionality is limited

[17:18] cherylj, I can later (~ 1.5 hours) === kadams54_ is now known as kadams54-away [17:47] since when do we have instance state on the status? [17:47] Bug #1522090 opened: Juju storage add fails with additional units [17:59] Bug #1522090 changed: Juju storage add fails with additional units [18:05] Bug #1522090 opened: Juju storage add fails with additional units === kadams54-away is now known as kadams54_ [18:08] Bug #1524021 opened: bootstrap timeout error should be more informative

[18:14] Bug #1524021 changed: bootstrap timeout error should be more informative

[18:17] Bug #1524021 opened: bootstrap timeout error should be more informative

[18:22] frobware, cherylj - you'll wanna see the last comment on bug 1523608 /o\ [18:22] Bug #1523608: Timed out waiting for device sys-subsystem-net-devices-juju-br0.device

[18:24] beisner: ah! glad you were able to get it to work. [18:24] and yeah, a better message would help [18:24] cherylj, yeah, can i get a refund on my 1.5 days? ;) [18:25] beisner: ha! if only I could refund time [18:27] ha! [18:28] cherylj, frobware so anyway - thanks for your work on those too. and future-work on the latter. [19:35] beisner, cherylj: looks like I return to peace and harmony [19:36] frobware: for now... [19:36] cherylj, You're gonna spoil xmas... [19:36] mwahahaha [20:07] frobware, cherylj - weeee! [20:21] alexisb: https://github.com/juju/juju/pull/3920 [20:29] katco: https://github.com/juju/retry/pull/3 [20:53] Bug #1524064 opened: provider/lxd: remove manual steps to add images

[22:06] Bug #1524077 opened: provider/azure: creating subnet fails if the environment name is too long

[22:09] Bug #1524077 changed: provider/azure: creating subnet fails if the environment name is too long

[22:23] dooferlad: http://juju-ci.vapour.ws/job/functional-container-networking/4/console [22:27] Bug #1524077 opened: provider/azure: creating subnet fails if the environment name is too long

[22:45] Bug #1524089 opened: update-status is rather noisy when there is nothing to do

[22:48] Bug #1524089 changed: update-status is rather noisy when there is nothing to do

[22:51] Bug #1524089 opened: update-status is rather noisy when there is nothing to do

[22:57] Bug #1524089 changed: update-status is rather noisy when there is nothing to do

[23:05] dimitern: want an easy code review? https://github.com/juju/juju/pull/3922 === arosales_ is now known as arosales [23:06] jam, looking [23:06] Bug #1524089 opened: update-status is rather noisy when there is nothing to do

[23:06] jam, ship it [23:09] jam, http://reviews.vapour.ws/r/3342/ [23:09] dooferlad, if you can spare some time I'd appreciate a look as well ^^ [23:09] mgz: review please :-) http://reviews.vapour.ws/r/3344/ [23:09] ack [23:33] dimitern, ping [23:34] dimitern, if you can drop me some email regarding the VLAN issue in the bridge script I'll take a look tomorrow. [23:37] * frobware__ EOD