[03:08] wallyworld: https://github.com/juju/juju/pull/7617 <- api dial timeout [03:09] looking [03:19] axw: lgtm, could you look at https://github.com/juju/juju/pull/7615 ? just a small tweak [03:19] wallyworld: yep, looking now [03:23] wallyworld: maybe a stupid question... just looking at the model-config code, isn't the answer to use "juju model-config --reset http-proxy" ? [03:24] axw: that will reset back to the model default, rather than explicitly "" [03:24] model default value may be something specific [03:24] ok [03:24] i'll check to be sure though [03:29] wallyworld: how's it going to work for string config attrs that won't accept an empty string value? it'll fail. you really need to delete in that case... but then the model defaults apply [03:30] axw: my take is that if the user tries to set "" and that is invalid, it will just error [03:30] i guess i could make foo= be nil [03:31] so we'd support foo="" -> empty string, foo= -> nil [03:32] wallyworld: the quotes will be interpreted by the shell [03:32] it'll end up as empty string in the args in either case [03:32] and that'd be pretty subtle anyway [03:32] right but I can check for "foo=" [03:33] hmmmm [03:33] wallyworld: anyway, it's an incremental improvement [03:33] yeah, we probably should have a separate semantic for delete [03:35] wallyworld: you've got some dupes in the test [03:35] all you need to add is "k1=", AFAICT [03:35] that was on purpose to reset the value back to non empty [03:35] ah [03:35] just to test parsing "" and empty [04:41] jam: did you have a few minutes to otuch base on the 2.2.2 bugs? [04:42] wallyworld: do you have a sensor on my chair? I *just* sat down :) [04:42] lol [04:42] i can wait [04:42] I was planning on looking through that list this morning, so I haven't really gotten to it yet [04:42] I'll have standup in 20 min [04:43] let me read through them and see what I can do [04:43] i had some info to share on my work [04:43] well, our work on this end [04:44] sure, I just figure if I haven't actually read through the list, a conversation may be a bit one sided :) [04:44] np, maybe ping me after standup then [04:46] wallyworld: so bug 1696739 and bug 1662857 both seem like they could just be removed from the 2.2.2 milestone [04:46] Bug #1696739: mongodb reporting saslStart and isMaster as being slow [04:46] Bug #1662857: cannot go get the source code [04:46] right [04:47] i wanted to create a new 2.2.3 milestonr [04:47] i changed the project driver to a team which we are in, but still can't do it it seems [04:48] also bug 1696557 [04:48] Bug #1696557: relationscopes may have an O(N^2) bug during startup [04:48] and i think bug 1697956 might fix bug 1696508 [04:48] Bug #1697956: restarting controller during an HA upgrade will cause it to not upgrade

[04:48] Bug #1696508: secondary controllers not coming up correctly

[04:49] and bug 1696472 can come off - we could do that for a 2.2.3 based on the history rpuning work [04:49] Bug #1696472: dblogpruner failed to prune logs by time [04:56] wallyworld: I created a 2.2.3, where did you try to create it? [04:56] I went to launchpad.net/juju and then clicked on the series 2.2, and then scrolled down for "create milestone" [04:56] i had thought there was a button on the milestone overview page [04:57] /shrug, maybe there should be, but I found it on the series [04:57] ah, yeah i see it there too [04:57] so i looked in the wrong place [04:59] jam: the biggest concern is that cdo qa blocker - the mongo issue. i was going to ping menno as he had had some involvement and could have maybe provided some insight but he's out today etc. i'm not sure what we can do about that bug before we look to cut a release [05:00] wallyworld: yeah, he handed some of that off to me, so I'll pick up from where he left off [05:00] there's 2 issues - the pinger one, and the other [05:00] controllers not being able to even login to mongo [05:01] bug 1701275 [05:01] Bug #1701275: [2.2.1] juju-agent loses connection to controller and doesn't retry to connect <4010>

[05:02] i think we still need to add the pinger one to the milestone as well [05:02] in addition to this other one [06:27] wallyworld: I'm back around again [06:27] finished my other meetings [06:38] wallyworld: so I'm actually focused on 1701294, which is a bug that is private, so I'm creating a public bug to track that against juju itself. [07:01] jam: sorry, was at school pickup. do you think we can release without fixing that other mongo one? I can't see us diagnosing, fixing, and testing this week, and we need 2.2.2 out the door === frankban|afk is now known as frankban [09:03] jam: if you have any free time today, I'd appreciate a review of https://github.com/juju/juju/pull/7606 (apiserver facade reorg) [09:03] yeah, trying to get through the backlog [09:03] thanks [09:03] I'm holding off on that one a little, as its bigger, but I'll try to get to it [09:03] no worries, no rush [09:07] axw: on pull 7611, you have a TODO about schema.Omit() [09:07] but from what I can tell your default is "", not schema.Omit() [09:09] * axw looks [09:10] jam: huh. I'm sure I had it as omit, must have changed it by mistake at some point. thanks, will fix - I believe the TODO still applies, but I'll double check [09:17] jam: thanks for catching: https://github.com/juju/juju/pull/7621. running QA now [10:26] jam: not sure I follow your comment about storageprovisioner - was that based off reading the description? I did put it under facades/agent, just didn't mention that in the description [10:27] axw: yeah, just reading the description [10:27] okey dokey [10:27] haven't gotten into the 385 files changed yet [10:27] :) [10:27] it's mostly just changed imports [10:27] 99% mechanical [10:27] axw: so is it mostly about asking if the grouping is sane? [10:28] jam: yup, which I think we agreed on already, but it needs the stamp [10:29] jam: actually the only non-mechanical bit in this PR are dropping the AuthClient check in apiserver/highavailability. there were auth fixes, but they were made in 2.2 and then merged already [12:09] wpk: axw: trivial review: https://github.com/juju/juju/pull/7624 [12:09] we had a helper function that was essentially "c.Assert(true)" [12:10] jam: doh :) lgtm [12:15] jam: whoops :) [12:20] /wind 54 [14:09] wpk: https://github.com/juju/juju/pull/7625 is another quick-tweak of logging data [14:13] wpk: I need to EOD now, but if you do a review I'm happy to land it [14:14] lgtm === frankban is now known as frankban|afk [21:56] externalreality: hey, could you please backport this PR into 2.2 branch? https://github.com/juju/juju/pull/7552 [21:57] OK [21:57] awesome, thanks [21:57] np [22:43] externalreality: bug 1696472 [22:43] Bug #1696472: dblogpruner failed to prune logs by time [22:44] nice, I'll make a leankit card now if one doesn't already exist