=== frankban|afk is now known as frankban === redir is now known as redir_sprint [09:24] Bug #1596462 opened: Deployment failed because state DB is locked

[09:27] Bug #1596462 changed: Deployment failed because state DB is locked

[09:42] Bug #1596045 changed: Juju says windows mongo: invalid version

[09:42] Bug #1596462 opened: Deployment failed because state DB is locked

[10:01] Bug #1596476 opened: juju charm resources does not map application name to charmstore url [10:46] Bug #1596493 opened: github.comjuju/juju/apiserver package: first record does not look like a TLS handshake

[10:55] Bug #1596493 changed: github.comjuju/juju/apiserver package: first record does not look like a TLS handshake

[10:58] Bug #1596493 opened: github.comjuju/juju/apiserver package: first record does not look like a TLS handshake

[11:10] Bug #1596496 opened: Bootstrap should fail early if '-d controller' is specified

=== urulama is now known as urulama-sprint [13:18] ey core, thanks for getting beta10 out the door Friday. *fires up the user-testing engine* [13:19] lazyPower: our pleasure :p [13:26] lazyPower: what's that engine btw? [13:27] dimitern my fingers and possibly duplicate bug reports ;) [13:27] * dimitern is interested in all sorts of testing automation we can share and reuse [13:27] ah :) [13:27] lazyPower: ok then, hope it goes fine :) [13:34] dimitern so far so good :) Yinzers do good work. [13:34] Bug #1576985 changed: aggregateSuite.TestBatching wrong size

[13:35] lazyPower: Yinzers ? [13:35] Thats Pittsburghese for "You all" [13:43] :) [13:43] Bug #1576985 opened: aggregateSuite.TestBatching wrong size

[13:52] Bug #1559400 changed: TestManageModelRunsRegisteredWorkers is flaky

[13:52] Bug #1576985 changed: aggregateSuite.TestBatching wrong size

[13:52] Bug #1580802 changed: NoContextWithLock fails on windows because of another process

[13:55] Bug #1559400 opened: TestManageModelRunsRegisteredWorkers is flaky

[13:56] Bug #1580802 opened: NoContextWithLock fails on windows because of another process

[14:05] Bug #1559400 changed: TestManageModelRunsRegisteredWorkers is flaky

[14:05] Bug #1580802 changed: NoContextWithLock fails on windows because of another process

[14:43] axw: ping [14:43] ericsnow: katco: hey, did you guys have time for a staus update hangout? [14:43] wallyworld: sure [14:44] ericsnow: katco: https://hangouts.google.com/hangouts/_/canonical.com/tanzanite-stand [14:53] Bug #1596559 opened: BootstrapSuite.TestRunTests

[15:03] katco: ping me when you're around [16:05] Bug #1596593 opened: juju show-model does not accept model as kwarg === frankban is now known as frankban|afk [16:22] ericsnow: i looked at the WIP forwarding worker. see if the comments make sense - i'm confused about what is expected as controller config; it seems it's been implemented as model config? also the worker loop seems a little unusually in not using a channel as the result of Next()? [16:22] wallyworld: thanks, I'll take a look [16:28] urulama-sprint: http://reviews.vapour.ws/r/5171/ [16:28] some juju reviewer plz: http://reviews.vapour.ws/r/5171/diff/# [16:29] frankban|afk: ^ PTAL [16:35] Bug # opened: 1596597, 1596603, 1596605, 1596607, 1596608, 1596609 [16:48] thumper: reviewed, shpit with a couple of gotchas [16:48] I must admit I did not go through the whoooole menn0 checklist but I covered many points [16:54] Bug #1596612 opened: show-controller output updates <2.0>

[16:54] Bug #1596615 opened: show-model output updates <2.0>

[16:54] Bug #1596616 opened: commands do not use two spaces between columns/headings <2.0>

[16:57] ericsnow: hey, i'm here. sorry, dentist appt. [16:57] katco: np [16:57] katco: just getting up to speed with wallyworld about status [16:58] katco: we need to get together at some point about how to integrate audit logs with log forwarding [16:58] * some point today [16:58] ericsnow: i'm free now if you'd like [16:59] katco: sure, moonstone? [16:59] ericsnow: brt [17:18] Bug #1596619 opened: remove aliases from juju commands <2.0>

[17:27] Bug #1596626 opened: juju gui not showing multiple models [17:36] Bug #1596626 changed: juju gui not showing multiple models [17:42] Bug #1596626 opened: juju gui not showing multiple models [17:54] Bug #1596626 changed: juju gui not showing multiple models [18:31] if we actally see cloud init apt-get failing suggesting update, why on earth would we not try that? [18:49] do we have some sort of wiki entry for the multiple mongo gotchas that got posted lately to the ML? [18:52] ericsnow: given all my changes, i think i need another proper review of: http://reviews.vapour.ws/r/5089/ [18:53] katco: k [18:53] ericsnow: ta [18:53] bogdanteleaga: not that i know of [18:54] ericsnow: i think i need to defer william's comments about RequestNotifier having internal knowledge of state, and tests for that type, but everything else is fair game [18:54] katco: k [19:54] Bug #1596687 opened: command list output not consistent <2.0>

[20:03] ericsnow: i'm ready to land http://reviews.vapour.ws/r/5089/. any thoughts? [20:03] Bug #1596687 changed: command list output not consistent <2.0>

[20:03] katco: one minute [20:06] ericsnow: np... took me 4 tries to land the previous one =| [20:12] Bug #1596687 opened: command list output not consistent <2.0>

[20:12] Bug #1596688 opened: normalizing the shares and user commands <2.0>

[20:17] katco: basically, ship-it [20:18] Bug #1596688 changed: normalizing the shares and user commands <2.0>

[20:19] ericsnow: ta [20:24] Bug #1596688 opened: normalizing the shares and user commands <2.0>

[21:07] ericsnow: are you familiar with the deploy api? [21:07] stokachu: a little [21:08] im trying to get the latest beta10 to deploy via the api [21:08] https://paste.ubuntu.com/17991715/ [21:08] we pass that to the apiserver but the response is always unable to find charm [21:09] was there an additional requirement added in beta10? [21:09] only thing we changed was to match the reworked api parameter keys === mwhudson_ is now known as mwhudson [21:25] stokachu: ok, sorry, i'm back and looking at this [21:25] thanks [21:26] ive been digging through the go code but nothing yet [21:26] katco: i did force a 'channel': 'stable' in the parameters [21:26] still failed though with same message [21:26] stokachu: hm [21:27] stokachu: do you happen to have the debug logs from the controller? [21:28] yea [21:29] what would you need? [21:32] stokachu: machine 0's log with debugging turned on and the deploy failure [21:32] stokachu: controller might give us a hint at what exactly is failing [21:32] whats the command to enable the debugging [21:33] stokachu: i always forget, sec. it's juju set-config log-level=DEBUG or something [21:34] juju set-config log-level=DEBUG [21:34] error: no application name specified [21:35] juju set-model-config logging-config='=DEBUG;unit=DEBUG' [21:40] stokachu: that looks right. easy, huh? where'd you find it? [21:40] juju help set-model-config [21:42] katco: http://paste.ubuntu.com/17994956/ [21:42] the bottom there has the latest deploy attempt [21:42] stokachu: ta [21:43] though it says params redacted [21:43] how do i get it to display it all [21:44] stokachu: probably can't; we hardcode that. let me 2xcheck though [21:45] stokachu: set level to trace [21:45] stokachu: and that should log the body as well [21:46] nice [21:47] katco: http://paste.ubuntu.com/17995148/ [21:47] thats everything [21:48] line 5204 shows the deploy [21:49] we don't pass a series as that usually is handled from the charmurl [21:50] ericsnow: feel free to take a peek too [21:53] stokachu: just on a whim, can you try passing the series? [21:53] sure [21:55] katco: same error [21:55] https://paste.ubuntu.com/17995455/ [21:55] 9203 [21:55] stokachu: k, worth a try [21:59] stokachu: i think this is where the error is coming from: https://github.com/juju/juju/blob/master/state/charm.go#L696 [21:59] stokachu: which would imply that it's not successfully being stored in mongo [22:00] weird [22:00] does the juju client code do something prior to calling the api? [22:01] as far as mongo is concerned [22:01] stokachu: yeah it does a few things, some related to resources [22:01] stokachu: that's all this: https://github.com/juju/juju/blob/master/cmd/juju/application/deploy.go#L301 [22:02] so did that change from beta9 to beta10? [22:04] stokachu: nope [22:06] :( [22:06] stokachu: https://github.com/juju/juju/commit/136d03f5987c89946b6987832c520c8879ea225f#diff-bad794df7c6f66424c0b9cc0961da6cc [22:06] stokachu: looks... plausible [22:07] so we have to upload first [22:07] wtf [22:09] stokachu: it... looks like that may be the case. it looks like that was a 1.16 guarantee... wow [22:09] so i've no idea where to go from here [22:09] do i need to start uploading every charm? [22:09] this PR looks like it has the $jfdi$ tag [22:10] lol with no review [22:11] stokachu: http://reviews.vapour.ws/r/5092/ [22:11] stokachu: unfortunately we are having to JFDI everything to meet deadlines, even though master is blocked [22:11] haha it just says shipit [22:12] stokachu: if there were no issues found, that's how the tool indicates it's safe to land [22:12] rick_h_: ping, you still up? [22:12] so im guessing deploying via the API with juju gui or another client is out of the scope of testing? [22:13] stokachu: upper management told us to move fast and break things as this is a beta [22:13] stokachu: alexisb usually bends over backwards to avoid doing this, but we lost that particular battle [22:15] stokachu: i don't know what the official stance on other clients is now. my guess is that yes, you have to upload charms/resources first... that seems weird to me [22:15] stokachu: seems like something the controller should be taking care of [22:15] yea [22:16] which it has been since forever [22:16] yeah [22:19] stokachu: can you send an email out to juju-dev asking for clarification about what's intended with a link to that PR? [22:19] katco: sure [22:19] stokachu: ta. sorry for the trouble. sometimes we don't have control over the experience :( [22:20] well the PR could've had more explanation at the very least [22:20] i understand its beta and we dont care about breaking things [22:55] ericsnow: wow, a lot of tests break if you add a new attribute to apiserver's ServerConfig struct :| [22:56] :( [22:56] ericsnow: i suspect it'll take me a bit just to hunt all these down (sigh). FULL STACK TESTING FTW! [22:56] katco: I feel your pain [22:57] ericsnow: do you? let's test it out. first, i need you to move to my city, then live in my house, etc. just to make sure we're exercising all aspects of me feeling pain [22:58] katco: heh