Juju doesn't retry or timeout container failures

[Adam Israel]

OS: Ubuntu 16.04.6
LXD: 3.12, installed via snap
Juju: 2.5.7-xenial-amd64, installed via snap
Cloud: lxd

Summary: If an lxd cloud fails to provide IP addresses for new containers, Juju will remain stuck in a allocating state when new charms are deployed.

I am/was encountering random dnsmasq failures. I was able to bootstrap a Juju controller to lxd. dnsmasq later failed, so when I tried to deploy a charm the unit was stuck in an 'allocating' state.

When I realised the problem, I restarted lxd, which restarted dnsmasq. The containers returned, this time with IP addresses, but the units remained in 'allocating' for quite some time. In the end I had to remove the applications or destroy the model.

Steps to reproduce.

1. Install LXD and Juju from snap
2. Configure LXD
3. Bootstrap a Juju controller to LXD
4. Kill the dnsmasq process
5. Deploy a charm
6. Wait for charm to enter 'allocating' state
7. `snap restart lxd`

What I expect to happen: Either the charm enter an error state because it can't be allocated, or for there to be a retry that continues once the network is available.

[Canonical Juju QA Bot (juju-qa-bot)]

This bug has not been updated in 2 years, so we're marking it Low importance. If you believe this is incorrect, please update the importance.