kelvinliuwallyworld: got this pr for the oci image fetching thing, could u take a look? thanks https://github.com/juju/juju/pull/1166200:33
wallyworldyup, looking00:34
wallyworldkelvinliu: can we log a message saying that microk8s was not ready if we reach the timeout?00:35
wallyworldalso, can we call a micrk8s command to see if the microk8s snap is still coming up so we can maybe know to wait a little longer00:36
kelvinliuwallyworld: it's because microk8s needs root to run00:36
kelvinliuno point to wait in this case00:37
wallyworldkelvinliu: are you saying that micrk8s k8s would have been broken even if juju were not there00:37
kelvinliuhttps://github.com/ubuntu/microk8s/blob/master/microk8s-resources/actions/common/utils.sh#L7 $USER wasnot there was the error00:38
wallyworldok. as an optimisation later, would be good for juju snap hook to check if user is in microk8s group at least so we avoid waiting for 30 unnecessarily00:40
kelvinliumicrok8s wasnot broken, it's just it's not accessible for non root user at that time00:40
kelvinliuwe are not waiting for 30s00:40
wallyworldah right it exits immediately00:40
kelvinliuwe  exit 0 immediately if we got any errors00:40
kelvinliuwe just exit immediately if we got any errors00:41
kelvinliunot only permission error00:41
wallyworldlgtm, with a suggestion for a message00:42
stubIs it a bug or by design that I can't deploy a charm with no hooks/*, just a dispatch script in $CHARM_DIR ?03:00
stubOr am I misinterpreting how dispatch works in 2.8 ?03:01
thumperstub: it is a bug03:14
thumpera bug that I thought we had caught already03:14
stubThink it was rc303:15
stubI can reproduce if you want03:15
thumperstub: if it is in rc3, it is in .003:19
stubthumper: reproduced with rc303:23
stubwant a bug?03:24
thumperstub: please03:25
mupBug #1881839: dispatch-only charms cannot be deployed, hooks/ directory required <juju:New> <https://launchpad.net/bugs/1881839>03:28
thumperhpidcock: if you get a moment, I'd like your thoughts on https://github.com/juju/juju/pull/1166003:37
hpidcockthumper: I'm not sure what tests were failing03:40
hpidcockwe really need to move to ecdsa03:40
hpidcockkey generation is fast enough + very secure03:41
hpidcockbut that is really only possible in 2.8 with tlm's changes to pki03:41
thumpera very simple bug fix https://github.com/juju/juju/pull/11665 for a race test failure04:25
=== arif-ali_ is now known as arif-ali
stickupkidmanadart, you around?09:40
manadartYeah, but I need 20 mins or so before I can do a HO.09:41
stickupkidmanadart, I need to nip out anyway for an early lunch, but the reason we don't land in the correct space with --bind is because we pick the wrong AZ09:42
stickupkidmanadart, in the StartInstanceParams AvailabilityZone the selection is wrong, just wanted to pick your brains about potentials of why09:43
stickupkidmanadart, the code doesn't look like it's changed from what I can see, so obviously missing something09:43
manadartstickupkid: But using constraints works. Mechanism for using bind should be the same. Anyway, I'll look and we can chat in a bit.09:44
stickupkidmanadart, that's what I would have thought, but clearly not09:45
stickupkidmanadart, ping11:18
manadartstickupkid: Pong.11:18
stickupkiddaily mc daily11:19
Eryn_1983_FLhey guys12:03
Eryn_1983_FLhow would i figure out why a node failed commission12:04
Eryn_1983_FLwhat/where are the logs12:04
stickupkidjuju status <- will tell you the over all issue12:07
stickupkidjuju debug-log --replay --no-tail <- will give you the logs for the model12:07
stickupkidjuju debug-log --replay --no-tail -m controller <- will give the logs for the controller12:07
stickupkidEryn_1983_FL, ^12:07
Eryn_1983_FLwell i think its cert issues and ovn issues,12:16
Eryn_1983_FLfun times..12:16
Eryn_1983_FLis Ubuntu 20.04 LTS good for openstack12:28
stickupkidmanadart, https://github.com/juju/juju/pull/1166715:16
stickupkidthat cleans up the INFAN networking15:16

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!