[00:04] <thumper> time for coffee, lunch, then tasks
[01:09] <wallyworld> thumper: you around?
[01:09] <thumper> yeah
[01:09] <wallyworld> thumper: can you join me and eric in a team standup ho?
[01:10] <thumper> yeah
[01:12] <axw> wallyworld thumper: so I can merge my storage stuff into develop now, yeah?
[01:12] <wallyworld> axw: yep
[01:12] <axw> sweet
[01:47] <thumper> :-(
[01:47] <thumper> I just tried 'juju login jaas'
[01:48] <thumper> that bit worked, but then...
[01:48] <thumper> $ juju models
[01:48] <thumper> ERROR cannot list models: unable to connect to API: x509: cannot validate certificate for 162.213.33.28 because it doesn't contain any IP SANs
[01:53] <thumper> how come this hasn't been tested before?
[01:53] <thumper> FFS
[01:57] <babbageclunk> thumper: Are you possibly using an IP address for a JAAS endpoint when others are using a DNS name?
[01:57] <thumper> babbageclunk: I just did 'juju login jaas' using the latest rc
[01:57] <thumper> I expect it should DTRT
[01:58] <babbageclunk> thumper: Ah, ok - no specific setup for the jaas controller then.
[01:58] <thumper> no
[01:59] <axw> thumper: are you up to date on develop? that was suppoed to be fixed by roger's latest PR
[02:01] <thumper> yes
[02:01] <thumper> perhaps his latest PR didn't land?
[02:01] <thumper> oh...
[02:01] <thumper> maybe I didn't merge
[02:01]  * thumper checks
[02:01] <thumper> no, I'm behind
[02:02]  * thumper updates and tries again
[02:04] <thumper> hah, works now
[02:04] <thumper> thank god
[02:04] <thumper> well, than rog
[02:04] <thumper> what's more, I have my first jaas model running
[02:04] <thumper> a canonical k8s model
[02:06] <rick_h> thumper: woot!
[02:07] <thumper> rick_h: I'm doing a meetup this afternoon
[02:07] <rick_h> thumper: nice, I got a talk accepted at pyohio
[02:07] <rick_h> thumper: so will be fun to plug some stuff
[02:10] <thumper> cool
[02:17] <wallyworld> babbageclunk: i made some suggested tweaks to the first pr, you ok for me to land?
[02:18] <axw> wallyworld: forgot to ask, did I miss anything in standup/yesterday?
[02:18] <wallyworld> axw: nah, same old, same old
[02:18] <wallyworld> all about the release
[02:18] <axw> wallyworld: cool. any news on the 2.1 hot fix?
[02:19] <babbageclunk> wallyworld: yup yup
[02:19] <wallyworld> tyvm for the review
[02:21] <babbageclunk> wallyworld: by the way, sorry not to roundtrip back to you about your review comments on https://github.com/juju/juju/pull/7441 - wanted to get it landed. Maybe have a look to see if the changes are what you meant?
[02:21] <wallyworld> sure, looking
[02:21] <wallyworld> babbageclunk: too late now anyway, i trusted whomever gave it a +1 :=)
[02:26] <babbageclunk> wallyworld: yeah, I figured that would be ok in the circumstances.
[02:27] <wallyworld> babbageclunk: we just have to hope people don't run out of disk space wen upgrading
[02:27] <babbageclunk> wallyworld: yeah, true.
[02:28] <babbageclunk> wallyworld: maybe we should reinstate the deletion - thumper, I know you were looking at performance when running the upgrade step, were you also looking at disk usage?
[02:33] <wallyworld> babbageclunk: ideally we'd run some tests on a system with the 4GB logs and see how much disk space is gobbled up during the upgrade
[02:39] <thumper> babbageclunk: yes I was looking at disk usage
[02:41] <babbageclunk> thumper: and the delete-as-we-go version didn't really help with that? Maybe we need to do something like compact the old collection as we go to get Mongo to release the space?
[02:48] <thumper> babbageclunk: not sure what speed impact that has
[02:48] <babbageclunk> thumper: hmm, true - probably not good.
[02:59]  * babbageclunk goes for a run
[05:24] <axw> wallyworld burton-aus: the 2.2.0 bump was made to develop, shouldn't that be on the 2.2 branch...?
[05:25] <wallyworld> axw: oh shit
[05:25] <wallyworld> yes
[05:25] <wallyworld> i'll fix develop
[05:25] <veebers> wallyworld, axw: ugh good catch, sorry that wasn't reflected in the release process doc.
[05:25] <axw> wallyworld: ok, thanks. just making sure I didn't send my changes in too early :)
[05:25] <axw> veebers: cool cool
[05:25] <wallyworld> nah, i didn't notice the merge target
[05:26] <veebers> wallyworld: does burton-aus need to re-do the pr?
[05:26] <wallyworld> yup
[05:29] <veebers> sweet he's on it
[05:29] <wallyworld> axw: this should fix develop https://github.com/juju/juju/pull/7453/
[05:29] <axw> wallyworld: thanks, LGTM
[05:30] <wallyworld> and merged
[16:52] <lazyPower> heyo juju dev, is there a way for me to "pick" which interface the unit will use for bootstrapping the controller? I'm having a jim dandy time trying to figure out why maas is reporting the management interface only during the bootstrap of the controller phase. the unit has 2 nic's available, both are online, one is addressable one is not, and it never seems to return that addressable interface.
[22:02] <wallyworld> externalreality: still in a call,running late
[22:17] <babbageclunk> thumper, wallyworld: created a PR against 2.2 for the TestPruneLogsBySize failure - should it have been against develop? https://github.com/juju/juju/pull/7458
[22:24] <wallyworld> babbageclunk: sorry, in call, but yeah, develop for now
[22:27] <babbageclunk> wallyworld: ok, I created it against 2.2 because it was in the 2.2.1 release blockers category
[22:27] <wallyworld> babbageclunk: that works too
[22:27] <wallyworld> better actually
[22:41] <babbageclunk> sweet
[22:44] <babbageclunk> Anyone not in that call who wants to do a trivial review for me? <bambi eyes> https://github.com/juju/juju/pull/7458
[23:24] <hml> babbageclunk: gh seems to be uphappy this evening.  :-/ i can’t see the diff
[23:24] <hml> babbageclunk: instead i’m getting angry unicorns
[23:24] <babbageclunk> hml: oh yeah - wow!
[23:25] <babbageclunk> that's pretty annoying
[23:25] <babbageclunk> hml: congratulations on the three months by the way!
[23:25] <hml> babbageclunk: ty!
[23:26] <babbageclunk> https://status.github.com/
[23:26] <hml> oops
[23:27] <babbageclunk> That's going to make a lot of things more difficult.
[23:30] <babbageclunk> hml: for reference, here it is as an old school patch: http://paste.ubuntu.com/24797094/
[23:31]  * hml looking
[23:34] <hml> babbageclunk: any idea why 5000 was chosen originally?
[23:35] <babbageclunk> hml: That's the threshold in the pruning code - it doesn't try to prune unless there are more than 5000 rows in the collection.
[23:36] <babbageclunk> hml: but obviously it wasn't a good choice in the test, because sometimes it'll end up with slightly more than 5000, getting pruned down to just under 5000.
[23:37] <hml> babbageclunk: that’s what I was just thinking…. the threshold doesn’t mean you can’t have less.
[23:39] <hml> babbageclunk: i’d approve if i could get to the correct page to do so
[23:39] <babbageclunk> hml: Thanks!
[23:40] <babbageclunk> hml: Ooh, the status just changed, and I can get to the PR now - try again?
[23:48] <babbageclunk> hml?
[23:49] <hml> babbageclunk: done
[23:49] <babbageclunk> hml: yaythanks
[23:50] <babbageclunk> hmm, not sure whether I'll be allowed to merge it though, since it's on the 2.2 branch.
[23:50] <balloons> babbageclunk, thank you for https://github.com/juju/juju/pull/7458
[23:51] <balloons> babbageclunk, i was just about to tell you about it. It's needed on develop too
[23:51] <babbageclunk> balloons: cool - I'll forward-port it now. Held up by the github outage!
[23:52] <thumper> externalreality: re bug 1696242, yes it is by design
[23:52] <mup> Bug #1696242: `juju debug log --help` shows both flags "--tail" and "--no-tail" as having a default value of `false`. <usability> <juju:New> <https://launchpad.net/bugs/1696242>
[23:53] <babbageclunk> balloons: Am I right that there's no mergebot watching 2.2? So I'll need to hassle someone to merge it?
[23:54] <externalreality> thumper, :-D
[23:54] <thumper> externalreality: I'm leaving  detailed comment
[23:54] <thumper> perhaps we should update the help
[23:54] <balloons> babbageclunk, yes. And veebers can help land it
[23:55] <veebers> babbageclunk: aye, Just need to figure out the parameters to fire the job off with
[23:57] <veebers> babbageclunk: was the one you need to land fix-prune-test