Instances with status error cannot be started again from the UI

[Simon Fels]

We nowadays allow starting instances which have the "error" status but asking the user to explicitly acknowledge AMS dropping all collected logs. In the UI starting an instance in the same way is currently not possible and need to be allowed.

The current CLI experience looks like this:

$ amc start cor2c1lkn8lfh8kbob10
Starting instance cor2c1lkn8lfh8kbob10 with status "error" will delete all collected logs. Do you want to proceeed [y/n]?

[David Edler (edlerd)]

Thanks for reporting the issue. Adding the option to start an instance in "error" status to the UI in this PR: https://github.com/canonical/anbox-cloud-dashboard/pull/799

[Michele Lo Russo (lorumic)]

@morphis I have reviewed and approved David's fix for the UI. You might want to reconsider how this is handled by the CLI though, as the current experience gives a postponed feedback for base instances:

ubuntu@juju-4d0777-prod-anbox-cloud-webdemo-2:~$ amc ls | grep cnk3rm8ta3a2f89u6rpg
| cnk3rm8ta3a2f89u6rpg | | test-vm | base | error | | lxd0 | 192.168.96.4 | | true | |
ubuntu@juju-4d0777-prod-anbox-cloud-webdemo-2:~$ amc start cnk3rm8ta3a2f89u6rpg
Starting instance cnk3rm8ta3a2f89u6rpg with status "error" will delete all collected logs. Do you want to proceeed [y/n]?y
Error: updating base instance not allowed

A better CLI experience would give "Error: updating base instance not allowed" before proposing the user to proceed with a start that is not really possible.