manual provider fails to remove itself if bootstrapping fails

[Dave Cheney]

The manual provider was unable to setup the machine correctly, something went wrong with mongodb so the jujud agent will eventually time out.

The manual provider failed to remove /etc/init/juju-db.conf because it got confused by the return code from pkill. Pkill returned 1 because it couldn't find any matching processes to signal, because that process had failed to start.

2014-03-24 01:02:01 ERROR juju.environs.manual bootstrap.go:102 bootstrapping failed, removing state file: rc: 1
2014-03-24 01:02:01 DEBUG juju.utils.ssh ssh.go:226 using OpenSSH ssh client
2014-03-24 01:02:01 DEBUG juju.utils.ssh ssh_openssh.go:147 running: ssh -o "StrictHostKeyChecking no" -o "PasswordAuthentication no" -i "/home/ubuntu/.juju/ssh/juju_id_rsa" "ubuntu@10.245.67.6" 'sudo pkill -6 jujud'
2014-03-24 01:02:06 ERROR juju.cmd.juju common.go:19 Bootstrap failed, and the environment could not be destroyed: rc: 1
2014-03-24 01:02:06 ERROR juju.cmd supercommand.go:296 rc: 1

[Andrew Wilkins (axwalk)]

Yep, the manual provider expects there to be a machine agent to take care of the environment destruction. We should (a) permit pkill to fail with rc 1, and have fallback code like in the local provider to handle failed bootstrap.

[Dave Cheney (dave-cheney)]

What about this sequence instead

$ [sudo] stop juju-db
$ [sudo] killall jujud
$ rm /etc/init/juju-* || true

On Mon, Mar 24, 2014 at 12:24 PM, Andrew Wilkins <
<email address hidden>> wrote:

> Yep, the manual provider expects there to be a machine agent to take
> care of the environment destruction. We should (a) permit pkill to fail
> with rc 1, and have fallback code like in the local provider to handle
> failed bootstrap.
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1296485
>
> Title:
> manual provider fails to remove itself if bootstrapping fails
>
> Status in juju-core:
> Triaged
>
> Bug description:
> The manual provider was unable to setup the machine correctly,
> something went wrong with mongodb so the jujud agent will eventually
> time out.
>
> The manual provider failed to remove /etc/init/juju-db.conf because it
> got confused by the return code from pkill. Pkill returned 1 because
> it couldn't find any matching processes to signal, because that
> process had failed to start.
>
> 2014-03-24 01:02:01 ERROR juju.environs.manual bootstrap.go:102
> bootstrapping failed, removing state file: rc: 1
> 2014-03-24 01:02:01 DEBUG juju.utils.ssh ssh.go:226 using OpenSSH ssh
> client
> 2014-03-24 01:02:01 DEBUG juju.utils.ssh ssh_openssh.go:147 running: ssh
> -o "StrictHostKeyChecking no" -o "PasswordAuthentication no" -i
> "/home/ubuntu/.juju/ssh/juju_id_rsa" "ubuntu@10.245.67.6" 'sudo pkill -6
> jujud'
> 2014-03-24 01:02:06 ERROR juju.cmd.juju common.go:19 Bootstrap failed,
> and the environment could not be destroyed: rc: 1
> 2014-03-24 01:02:06 ERROR juju.cmd supercommand.go:296 rc: 1
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/juju-core/+bug/1296485/+subscriptions
>