wallyworld | babbageclunk: it seems like that maas test is running but just takes a long time, comparing the output of a successful run to the failed one. maybe just increase the timeout, and take a look a finfolk to see if stuff needs cleaning up etc | 00:54 |
---|---|---|
babbageclunk | wallyworld: yeah, will do | 00:54 |
timClicks | is this the current release template? https://discourse.jujucharms.com/t/juju-release-template/1506 | 01:13 |
timClicks | I would like to add a check that if a config (app/model) is added, that the docs are updated to suit | 01:14 |
wallyworld | timClicks: release template is now a google doc | 01:21 |
wallyworld | on the shared drive | 01:21 |
timClicks | that's what I thought, which is why I asked | 01:21 |
wallyworld | thumper: there's a race in the model cache, got a minute to talk about it? | 01:21 |
thumper | wallyworld: yep | 01:21 |
wallyworld | let's go to 1:1 | 01:21 |
thumper | ack | 01:22 |
thumper | wallyworld: https://github.com/juju/juju/pull/11171 | 01:57 |
thumper | wallyworld: what does the names package use from utils? | 02:00 |
wallyworld | i'd need to check again, one sec | 02:00 |
wallyworld | thumper: IsValidUUIDString() for one | 02:01 |
thumper | ugh | 02:02 |
thumper | I've been quietly trying to kill the utils package for some time | 02:02 |
wallyworld | also, MustNewUUID() | 02:02 |
thumper | make smaller targeted packages that fit | 02:02 |
wallyworld | i think heather is doing some work in that area | 02:03 |
thumper | perhaps we should make a juju/uuid package for those | 02:03 |
wallyworld | +1 | 02:03 |
wallyworld | heather has a PR up to take another step | 02:03 |
thumper | wallyworld, babbageclunk: https://github.com/juju/juju/pull/11171 | 02:42 |
thumper | or anyone really... | 02:42 |
babbageclunk | thumper: looking | 02:42 |
hpidcock | wallyworld: did you say you did some snap work already for go1.13? | 02:43 |
wallyworld | hpidcock: there's a PR with my juju snapcraft changes | 02:44 |
wallyworld | https://github.com/juju/juju/pull/11138 | 02:45 |
hpidcock | wallyworld: thx | 02:45 |
wallyworld | tlm: do you think we'll get the watcher stuff landed today? | 04:41 |
tlm | wallyworld, hpidcock HO for some rubber ducking ? | 04:42 |
wallyworld | tlm: just finishing a meeting, can be there soon | 04:42 |
tlm | k | 04:42 |
tlm | wallyworld: nvm hpidcock help me sort it out. | 04:56 |
tlm | I think I am on the last failing test now so should be soonish | 04:56 |
wallyworld | tlm: gr8 ok, sorry still in meeting | 04:56 |
tlm | np | 04:57 |
wallyworld | tlm: still all good? | 05:22 |
tlm | wallyworld: yep, giving it a full run | 05:23 |
tlm | sent DM | 05:23 |
wallyworld | all good, i can review when ready | 05:23 |
hpidcock | wallyworld: can you remember if you ever did a remote-build using your changes? | 06:32 |
hpidcock | it's taking an eternity to upload to lp for me. Not sure if you had a similar experience? | 06:33 |
tlm | are you sure it's still working? I had similar issues but it had failed and gone into an endless loop | 06:36 |
tlm | not saying this is the case now hpidcock, just an FYI | 06:36 |
hpidcock | could just be slow upload | 06:38 |
hpidcock | i dunno the git repo has nothing in it | 06:38 |
tlm | restart your computer 3 times ? | 06:41 |
hpidcock | Three shall be the number thou shalt count, and the number of the counting shall be three. | 06:43 |
tlm | the number of pringles in a stack | 06:44 |
wallyworld | hpidcock: i did do a remote build to test it | 06:45 |
wallyworld | it took a little while but to too long | 06:45 |
hpidcock | must be my Australian internet | 06:46 |
wallyworld | maybe try running on a jenkins slave? | 06:46 |
hpidcock | alright for some reason, it is uploading 10gb | 06:48 |
hpidcock | that might be my problem | 06:48 |
wallyworld | wow, i don't recall that much but didn't check too closely | 06:48 |
wallyworld | i would hope it's not doing the entire git repo | 06:49 |
hpidcock | I think it has pulled in my .git folder | 06:49 |
wallyworld | it just would need the source files | 06:49 |
wallyworld | if it's pulled in .git i'd raise a bug and see what sergio says | 06:50 |
wallyworld | we can avoid that in the build scripts | 06:51 |
tlm | wallyworld, hpidcock PR pushed, no doubt it may have some problems | 06:51 |
wallyworld | looking | 06:51 |
wallyworld | tlm: you'll need to address the existing comments | 06:53 |
tlm | already on it :) | 06:53 |
wallyworld | tlm: there's still a bunch of import grouping issues | 07:13 |
tlm | yeah I put a comment on one of them | 07:13 |
tlm | not sure what you where chasing v what is there ? | 07:13 |
wallyworld | there's 3 blocks we use: std lib, 3rd party libs, juju/juju | 07:14 |
wallyworld | so see operator.go - it has juju/juju stuff mixed in with k8s.io stuff and juju /errors etc | 07:15 |
wallyworld | same with namespaces.go | 07:15 |
tlm | so we consider juju/errors external for this purpose ? | 07:18 |
wallyworld | yup | 07:19 |
tlm | ah no worries makes sense now | 07:19 |
wallyworld | anything not from juju/juju is 3rd party | 07:19 |
wallyworld | confusing for sure | 07:19 |
wallyworld | tlm: to allow parallel, work, i submitted a couple of comments - main thing is we already have a mock notify watcher that can be used | 07:20 |
wallyworld | so no need to create a new one for k8s testing | 07:20 |
wallyworld | i inlcuded a reference to the code in the comment | 07:21 |
tlm | ta looking now | 07:21 |
wallyworld | tlm: i'm trying tp get my head around the new k8sWatcherFn on the test suite. do we need it instead of what was there before? also, there's maybe one snafu in there? | 07:32 |
wallyworld | if s.k8sWatcherFn != nil { | 07:32 |
wallyworld | w, err = s.k8sWatcherFn(informer, name, clock) | 07:32 |
wallyworld | } | 07:32 |
wallyworld | w, err = provider.NewKubernetesNotifyWatcher(informer, name, clock) | 07:32 |
wallyworld | ... | 07:32 |
wallyworld | i think we can land this tomorrow and do the stringswatcher also | 07:33 |
tlm | I don't mind, almost got it swapped out for the notify one | 07:33 |
wallyworld | ok | 07:34 |
tlm | want me to HO to explain the above ? | 07:34 |
wallyworld | sure | 07:34 |
nammn_de | manadart: if you have some time later, beside it being ci day, want to go over 2 of your comments on this pr? https://github.com/juju/juju/pull/11143 As I want to make sure to understand what is meant | 09:25 |
manadart | nammn_de: OK, if not today, then tomorrow. | 09:48 |
nammn_de | manadart: sure, just give me a ping if you can | 09:49 |
nammn_de | stickupkid: https://github.com/juju/juju/pull/11173 regards running focal with force | 10:21 |
stickupkid | nammn_de, i think this needs to target 2.7 | 10:23 |
nammn_de | stickupkid: oh yes | 10:23 |
nammn_de | stickupkid: https://github.com/juju/juju/pull/11174 | 10:31 |
stickupkid | nammn_de, lovely | 10:33 |
nammn_de | stickupkid: ta! | 10:34 |
nammn_de | wow it feels like we didnt forward port 2.7 to develop for ages. I am currently processing all those merge conflicts.. | 12:49 |
skay | I have a newbie postgresql charm question. by default postgres charm stores backups in /var/lib/postgresql/backups. when using storage, pgdata is /var/lib/postgresql/10/main. is it safe to store the backups folder there instead so that it is saved in the attached storage? | 13:41 |
skay | when I've looked at how someone else is configuring this, they just used the defaults | 13:42 |
nammn_de | manadart: https://github.com/juju/juju/pull/11177 the one I closed and already resolved mostly. Maybe of help to you. | 13:56 |
stub | skay: Better to put your backups dir in /srv/pgdata. /srv/pgdata is the Juju storage mount. /var/lib/postgresql/10/main should be a symlink to /srv/pgdata/10/main IIRC | 15:02 |
manadart | nammn_de: Superseding patch: https://github.com/juju/juju/pull/11178 | 15:20 |
nammn_de | manadart: approved! | 15:26 |
manadart | nammn_de: Thanks. | 15:26 |
skay | stub: ack | 15:33 |
skay | stub: if I understand the source, when storage is detached everything is destroyed. at that point the unit is probably being destroyed so it doesn't matter. do I understand that correctly? | 15:39 |
skay | I'm just really paranoid about backups and want to be certain | 15:40 |
skay | (by doesn't matter, I mean, I know the destruction is going on and I have backups saved elsewhere) | 15:40 |
nammn_de | stickupkid: https://github.com/juju/juju/pull/11179 | 16:34 |
nammn_de | stickupkid: tested it all, now bootstraps fine for me locally. | 16:35 |
stickupkid | nammn_de, wicked | 16:35 |
stickupkid | approved | 16:35 |
nammn_de | stickupkid: ta! | 16:36 |
hml | rick_h: any issues with canonical irc? | 17:08 |
hml | stickupkid: i’m thinking microk8s won’t work in a lxc container. i get weird results like the enable storage works, but it never reaches a running status | 17:09 |
hml | stickupkid: i can also start a pod that never completes | 17:10 |
nammn_de | stickupkid and are having issues with irc as well | 17:20 |
nammn_de | *s/are/I | 17:20 |
stickupkid | nammn_de, i reconnected and it's fine now | 17:21 |
stickupkid | nammn_de, it failed | 17:56 |
stickupkid | nammn_de, we need to work out why that failed | 17:56 |
nammn_de | stickupkid: meh, I am so sad. Maybe I did the lxd configuration wrong? | 17:57 |
stickupkid | rick_h, first attempt bootstrapping focal on our CI, it works locally, but fails in CI | 17:58 |
stickupkid | rick_h, checking out why now | 17:58 |
nammn_de | but I just followed the instruction.. I need to leave now. Gonna look at this later or next week. | 17:59 |
nammn_de | stickupkid: if you have some updates I am more than happy to read them =D | 17:59 |
stickupkid | nammn_de, sure | 17:59 |
stickupkid | ah, so it can't ssh in | 17:59 |
stickupkid | rick_h, you got a sec? | 18:02 |
rick_h | stickupkid: sure thing if you're free | 18:13 |
stickupkid | daily | 18:13 |
rick_h | omw | 18:13 |
thumper | morning team | 19:34 |
hml | morning thumper | 19:34 |
rick_h | morning thumper | 19:36 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!