debug-log does not terminate when the environment is destroyed

Bug #853920 reported by Roger Peppe
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
pyjuju
Triaged
Low
Unassigned

Bug Description

When running juju debug-log, I would expect it to finish when the juju environment
is destroyed. Instead, I see something like the following output, and then nothing more.

2011-09-19 15:36:50,907 ERROR SSH forwarding error: Connection to ec2-50-16-111-175.compute-1.amazonaws.com closed by remote host.

Unhandled error in Deferred:
Traceback (most recent call last):
Failure: juju.errors.NoConnection: SSH forwarding error: Connection to ec2-50-16-111-175.compute-1.amazonaws.com closed by remote host.

Revision history for this message
Kapil Thangavelu (hazmat) wrote : Re: [Bug 853920] [NEW] debug-log does not terminate when the environment is destroyed

I just wanted to explain why this behavior exists.

Right now destroy-environment is entirely client-side to ensure that its robust,
regardless of the state of the environment. Ie. any number of critical
infrastructure pieces could be dead, and destroy-environment will still work.

This of course makes coordination around environment destruction difficult.

Excerpts from Roger Peppe's message of Mon Sep 19 14:44:11 UTC 2011:
> Public bug reported:
>
> When running juju debug-log, I would expect it to finish when the juju environment
> is destroyed. Instead, I see something like the following output, and then nothing more.
>
> 2011-09-19 15:36:50,907 ERROR SSH forwarding error: Connection to
> ec2-50-16-111-175.compute-1.amazonaws.com closed by remote host.
>
> Unhandled error in Deferred:
> Traceback (most recent call last):
> Failure: juju.errors.NoConnection: SSH forwarding error: Connection to ec2-50-16-111-175.compute-1.amazonaws.com closed by remote host.
>
> ** Affects: juju
> Importance: Undecided
> Status: New
>

Revision history for this message
Clint Byrum (clint-fewbar) wrote :

Excerpts from Kapil Thangavelu's message of Mon Sep 19 16:00:23 UTC 2011:
> I just wanted to explain why this behavior exists.
>
> Right now destroy-environment is entirely client-side to ensure that its robust,
> regardless of the state of the environment. Ie. any number of critical
> infrastructure pieces could be dead, and destroy-environment will still work.
>
> This of course makes coordination around environment destruction
> difficult.

That makes perfect sense.

We could have a best-effort try to set something in zookeeper that says
basically "environment is doomed" so they can gracefully give up if they
happen to get that message, right?

I've experienced this before, so setting status to Confirmed.

This being an annoyance and not really breaking any functionality,
setting importance to Low,

Changed in juju:
status: New → Confirmed
importance: Undecided → Low
Curtis Hovey (sinzui)
Changed in juju:
status: Confirmed → Triaged
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.