Machine agent hangs if lxc container start fails

[Kapil Thangavelu]

Reported by nick barcet on irc. The machine agent will hang on lxc-wait -s RUNNING if the container start fails.

[Kapil Thangavelu (hazmat)]

The machine agent should wait on the RUNNING|STOPPED, and log the error, and possibly adjust the status on the unit. Understanding why the container fails to start correctly is a separate issue, but useful to discuss in the context of this bug report for now.

[Kapil Thangavelu (hazmat)]

Attached, container console log from a failed attempt at starting the container. It looks like upstart just exits.

[James Westby (james-w)]

Duplicate of bug 907655?

[Kapil Thangavelu (hazmat)]

lxc has upgraded its implementation of wait to be re-entrant (ie can issue
multiple lxc-wait), so this may no longer be an issue, albeit there are
several conditions that need to be bitwise-or'd to capture the full set.
Looking over the golxc implementation, it looks like it only waits on the
same start state, so a failure to start there will cause the same issue of
a goroutine stall.

On Fri, Oct 11, 2013 at 11:42 PM, Curtis Hovey <email address hidden> wrote:

> ** Changed in: juju
> Status: Confirmed => Triaged
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/912879
>
> Title:
> Machine agent hangs if lxc container start fails
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/juju/+bug/912879/+subscriptions
>