=== frankban|afk is now known as frankban === rumble is now known as grumble [15:08] Have there been any notable changes to `juju debug-log`? If I run it with --tail, it will sometimes exit on its own. [15:38] aisrael: i've noticed that too, but it only seems to happen when tailing a log from a jaas controller. is that the same for you? [15:39] might also be that debug-log uses an ssh connection that shuts down after x mins of inactivity based on client config. i haven't dug too deeply into it yet. [15:39] ^^ speculating on the ssh bits there.. i really don't know how debug-log gets its data [15:45] kwmonroe: Nope, this is a local controller. I haven't poked around the controller machine's logs yet, but that'll probably be the next step === hml_ is now known as hml [16:38] hi here, I have my kubernetes-worker in "Error", and when I looked at "juju debug-log", I saw this: [16:38] unit-kubernetes-worker-1: 16:37:21 ERROR juju.worker.dependency "metric-collect" manifold worker returned unexpected error: failed to read charm from: /var/lib/juju/agents/unit-kubernetes-worker-1/charm: open /var/lib/juju/agents/unit-kubernetes-worker-1/charm/metadata.yaml: too many open files [16:38] I can "cat" this file on the kubernetes-worker-1 unit [16:39] so I don't understand why it returns a "too many open files" :( === frankban is now known as frankban|afk [18:28] Zic: you running on lxd? i've seen "too many open files" on large(ish) lxd deployments and have fixed them with this guide: https://github.com/lxc/lxd/blob/master/doc/production-setup.md [18:28] specifically the limits/sysctl changes ^^