ERROR juju.worker.meterstatus error running "meter-status-changed": charm missing from disk

[John A Meinel]

As part of just doing a normal deploy of a charm, I ended up getting:
unit-ubuntu-operator-4: 08:53:14 ERROR juju.worker.meterstatus error running "meter-status-changed": charm missing from disk

We shouldn't have an ERROR in a common workflow.
First, because we are removing things like budget, etc, we can probably get rid of meter-status-changed in 3.0. But even further we should never trigger an error. Whatever the cause of the above, it sounds like we are trying to run meter-status-changed before we've even unpacked the charm on disk.

[Alok Prudhvi (alokprudhvi)]

I'm facing the same issue on mysql-innodb-cluster and i see the unit is in blocked state.

[Jordan Barrett (barrettj12)]

I'm experiencing the same error message in Juju 3.0-beta4-af9f362 when bootstrapping an LXD controller (error from the controller charm):
```
unit-controller-0: 15:18:06 ERROR juju.worker.meterstatus error running "meter-status-changed": charm missing from disk
```

SSH'ing into the machine, the charm directory is still intact, so not sure what's happening.

[Alok Prudhvi (alokprudhvi)]

What i've identified in my environment is the core20 snap is not allowing mysql to start and the see few errors messages on that. So tried rebooting the lxd container and now i complete lost the control of the unit and i logged into host and tried to do an lxc list which is also not working and then saw the snap list and found that the same core20 snap is in disable state and don't know what's the issue. I'm still stuck with this problem and not able to recover the mysql-cluster

[Alex Lutay (taurus)]

The error message is constantly reproducible on juju 3.1.7 with lxd "5.0.3-ffb17cf (27037) 5.0/stable" and juju 3.3.1 (snap revision 25912 from 3/stable).

STR:

> multipass launch -c 12 -m 32G -d 100G --timeout 100500 -n juju317 charm-dev
> multipass shell juju317
> juju switch lxd && juju add-model test
> juju deploy postgresql # the same issue for all Data charms, e.g. mysql, s3-integrator, ...

Actual results (the list of errors including one in $subject):

> ubuntu@juju317:~$ juju debug-log --tail | grep ERROR
> machine-0: 09:41:35 ERROR juju.worker.dependency "broker-tracker" manifold worker returned unexpected error: no container types determined
> machine-0: 09:41:35 ERROR juju.worker.dependency "kvm-container-provisioner" manifold worker returned unexpected error: container types not yet available
> machine-0: 09:41:35 ERROR juju.worker.dependency "lxd-container-provisioner" manifold worker returned unexpected error: container types not yet available
> unit-postgresql-0: 09:41:35 ERROR juju.worker.meterstatus error running "meter-status-changed": charm missing from disk

We do not see any implications of this ERROR message on the charm functionality, so JFYI only here.

Note: the same error message shown for "juju add-unit postgresql".
P.S. Juju on (Micro)K8s has no such error messages.