[00:26] <thumper[m]> added
[00:27] <tlm> matrix thumper
[00:27] <thumper[m]> I've added a #juju channel in matrix.org with an IRC bridge
[00:28] <timclicks[m]> removed it from the juju-devops room I created
[00:28] <thumper[m]> Just messing around testing just now
[00:31] <tlm[m]> howdy
[00:40] <thumper> it looks like the matrix bridge only creates users on demand...
[00:41] <timClicks> I think we confused the bot because there were two matrix rooms bridged
[00:43] <thumper> have you removed the second one?
[00:47] <thumper[m]> https://github.com/juju/juju/pull/11313 up for review
[00:47] <TimM[m]> <thumper "have you removed the second one?"> yes, that room is now empty
[00:48] <tlm[m]> thats working
[00:48] <TimM[m]> lol yup
[00:48] <thumper[m]> uh ha, I see many new folks
[00:48] <thumper[m]> also, I was curious as to how the reply would look in irc
[00:49] <TimM[m]> pretty keen to charm up a matrix home server now
[00:49] <tlm[m]> was having the same thought
[00:49] <thumper[m]> I think ec0 has done that
[00:49] <thumper[m]> or at least some aspects of that
[00:50] <TimM[m]> https://github.com/alchemy-charmers/charm-matrix
[00:52] <ec0[m]> yep, that one :)
[00:52] <ec0[m]> Tim M:
[00:52] <ec0[m]>  * Tim M: let me know if you get stuck on it
[00:53] <ec0[m]> there's still some rough edges but the snap stuff has been merged upstream, and the charm does deploy a working homeserver
[00:53] <TimM[m]> I was thinking about Juju for web workloads, and discovered that the latest series that fail2ban supports is trusty
[00:54] <TimM[m]> I made several edits yesterday after work and I have it working for more recent series also
[00:54] <TimM[m]> it's still based on charmhelpers, but that can change too over time
[01:00] <wallyworld> timClicks: just saw this "Controller independent of the k8s cluster: A Juju controller must sit outside of Kubernetes."
[01:00] <wallyworld> in the mapping juju concepts
[01:00] <timClicks> Oh right, yes I meant to clarify that
[01:00] <wallyworld> that's no longer true
[01:00] <wallyworld> used to be early on
[01:00] <timClicks> Cool, I thought so, will update
[01:00] <wallyworld> ty
[01:19] <tlm[m]> wallyworld kelvinliu pushed up the tmp work around
[01:21] <kelvinliu> yup, will have a look, ty
[02:13] <thumper> wallyworld: this is the PR we talked about https://github.com/juju/juju/pull/11313
[02:13] <wallyworld> sorry, getting to it, will look now
[02:13] <thumper> ack
[02:15] <wallyworld> thumper: lgtm. love deleting tests from state
[02:15] <thumper> thanks
[02:15] <wallyworld> if we delete enough, maybe no more timeouts
[02:15] <thumper> if we move the check tests to github actions, we may have the same
[02:15] <wallyworld> indeed
[06:38] <TimM[m]> ec0:  here is the fail2ban charm I was talking about https://github.com/timClicks/fail2ban-charm
[10:08] <stickupkid> manadart, did you land the model cache changes in 2.7 or 2.8 ?
[10:08] <stickupkid> manadart, https://mail.google.com/mail/u/0/#inbox/FMfcgxwGDNVHpPHvPqQWhcTLDkdQLjNz
[10:09] <stickupkid> manadart, https://bugs.launchpad.net/juju/+bug/1860970
[10:09] <mup> Bug #1860970: SetCharmProfiles is called too frequently <juju:Triaged> <https://launchpad.net/bugs/1860970>
[11:11] <stickupkid> manadart, wtf is this https://github.com/juju/juju/pull/11279#issuecomment-598669584 ?
[11:13] <manadart> ¯\(°_o)/¯
[11:13] <stickupkid> silly bot
[11:42] <manadart> stickupkid: Got a sec for https://github.com/juju/juju/pull/11316 ?
[11:44] <stickupkid> manadart, done
[11:44] <manadart> stickupkid: Ta.
[12:41] <hml> stickupkid:  why does https://github.com/go-goose/goose/pull/82 include the deprecated commit as well?  got really confused in review.  :-D
[12:42] <rick_h> sorry achilleasa looking
[12:42] <stickupkid> hml, I wanted to rebase once deprecated one has landed
[12:42] <stickupkid> :)
[12:42] <achilleasa> rick_h_: so I am nearly there with the relation-created hook https://pastebin.canonical.com/p/CJkg76rpKS/ I am still trying to get rid of that spurious relation-departed hook
[12:43] <achilleasa> is this the right order of hooks? install, xxx-relation-created, leader-elected, ... ?
[12:43] <rick_h> achilleasa:  yea, so after the install (deps/etc are available on the system) and before leader-elected. So that it can check if there are/will be other units coming in.
[12:43] <stickupkid> rick_h, achilleasa manadart https://github.com/juju/juju/pull/11317
[12:44] <achilleasa> rick_h: and this is how it looks for the non-leader unit: https://paste.ubuntu.com/p/3vGczmCQ7v/
[12:45] <stickupkid> rick_h, exactly, doing that step now
[12:46] <rick_h> achilleasa:  cool!
[12:46] <stickupkid> rick_h, offically /usr/bin/env python should work, but it doesn't - shame
[12:46] <rick_h> stickupkid:  well it's not "python" it's python3
[12:46] <rick_h> that works
[12:46] <rick_h> the thing is that if you want it to work you'd have to symlink python3 to just python
[12:46] <rick_h> it's the "official way"
[12:47] <stickupkid> :(
[12:47] <rick_h> stickupkid:  for our purposes, given we know the system, we should just prefix the .py script command in the dh with the right python for the system
[12:48] <rick_h> stickupkid:  e.g. HOME=obj-x86_64-linux-gnu/home /usr/bin/$PYTHON /<<PKGBUILDDIR>>/src/github.com/juju/juju/scripts/generate-docs.py man -o obj-x86_64-linux-gnu/juju.1
[12:48] <rick_h> where $PYTHON is figured out with shell foo for "should it be path to py2 or py3"
[13:11] <stickupkid> rick_h, yeah, works for 2.7 and 3 https://github.com/juju/juju/pull/11317/files
[14:04] <hml> stickupkid: approved 81
[14:04] <hml> looking at 82
[14:12] <achilleasa> hml: can you take a look at https://github.com/juju/charm/pull/305 ?
[14:13] <hml> achilleasa:  will do after current review in progress.  :-)
[14:13] <achilleasa> hml: not in a hurry
[14:17] <hml> stickupkid: 82 is approved with some suggestions.
[14:28] <manadart> Bionic packages for 2.7.4 are in stable. Confirmed and tested. Working on the others.
[14:40] <hml> achilleasa:  approved
[14:55] <manadart> Eoan is there.
[15:52] <admcleod> having a problem fetching interfaces, e.g. https://juju.github.io/layer-index/interfaces/haclster.json
[15:52] <admcleod> feels like a faulty node in a pool somewhere
[15:52] <admcleod> or a firewall rule here somewhere
[19:24] <hml> admcleod: looks like the whole site is down.  not sure who is around right now to help.  maybe an RT?  grasping here.  :-)