ballot | jam: following our discussion yesterday about podspec "spec" documentation : https://bugs.launchpad.net/juju/+bug/1889519 | 08:26 |
---|---|---|
jam | thanks ballot | 08:26 |
barryprice | jam: oh one other thing, I noticed that on deploy, my units get Workload set to 'active' in juju status, but after a config change, new units are spawned and their Workload becomes 'unknown' | 09:33 |
barryprice | not very high priority but I thought I was handling that | 09:33 |
jam | barryprice, let me look | 09:34 |
mthaddon | barryprice, jam: I think that's because in configure_pod you're only setting status on the leader? | 09:34 |
barryprice | mthaddon: it affects the leader too though | 09:35 |
barryprice | hmm or maybe the old leader sets spec which causes new pods to come up and until config-changed next fires, there's no status set? | 09:35 |
barryprice | I'm a bit fuzzy about the order in which things happen there, do the new pods come up during config-changed or after it's fired? | 09:36 |
jam | barryprice, so none of the other units have a status set (as tom noted). For the leader itself, I would think that Juju would look at the readiness probe to determine when the pod has transitioned to fully ready | 09:36 |
jam | after being restarted | 09:36 |
jam | barryprice, IIRC, you should set the status to active, and then Juju will handle the fact that it isn't actually active while the pod is restarting. | 09:37 |
barryprice | ha, ok - thanks. this is low pri, it was just bugging me | 09:37 |
mthaddon | I'd expect the pods are rescheduled if there's a config change by "self.model.pod.set_spec(spec)" | 09:37 |
mthaddon | jam: https://paste.ubuntu.com/p/7n6t8kbq6M/ does that look like a reasonable .jujuignore for the jenkins-agent charm? | 10:18 |
jam | mthaddon, looks reasonable to me. Juju has some default ignores (like build), but I think it is fine to be explicit in your rules | 10:19 |
mthaddon | ack - I wasn't sure if there were defaults, but sounds good, thx | 10:19 |
barryprice | mthaddon: fyi https://github.com/canonical/charmcraft/blob/master/charmcraft/jujuignore.py#L208 | 10:33 |
mthaddon | ah thx | 10:34 |
mthaddon | looks like I need to include trailing / for directories, will update | 10:36 |
facubatista | :hlo: | 11:14 |
facubatista | ¡Muy buenos días a todos! | 11:14 |
facubatista | jam, did you see my mail about release notes? I'd try to do the release now (in case my electricity provider decides to ruin my day later) | 11:17 |
jam | facubatista, morning. I did see them, I'll look over them once more and confirm. They seemed ok | 11:17 |
facubatista | jam, thanks! | 11:18 |
jam | facubatista, meeting? | 12:02 |
facubatista | oops | 12:03 |
mup | Issue operator#249 closed: Let's instrument all registered methods' calls <bug> <Created by facundobatista> <Closed by jameinel> <https://github.com/canonical/operator/issues/249> | 12:23 |
ballot | Hello there. Is there a way to retrieve the config value set for juju-external-hostname config item ? I'm logging self.model.config and the key doesn't seem to be here even when set. | 12:50 |
* facubatista -> errands, bb~1h | 15:20 | |
* facubatista eods | 21:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!