Hung agent processes

Bug #1043292 reported by Adrien
36
This bug affects 7 people
Affects Status Importance Assigned to Milestone
puppet (Ubuntu)
Confirmed
High
Unassigned

Bug Description

Since 2.7.11, the puppet process has gotten stuck in some hosts.

You can read the issue on the bugreport below:

http://projects.puppetlabs.com/issues/13038

Is it possible to update the package to 2.7.12 or > in LTS?

Thx.

Revision history for this message
James Page (james-page) wrote :

Thanks for taking the time to report this bug in Ubuntu.

Unfortunately its not possible to upgrade the version of puppet in Ubuntu 12.04; if its possible to identify what was causing this issue and what the fix in puppet was this could be fixed through the Stable Release Update process.

Changed in puppet (Ubuntu):
importance: Undecided → Medium
Revision history for this message
Clint Byrum (clint-fewbar) wrote :

I have seen this as well on 12.04 with just simple "puppet apply".

Looking through 2.7.12's changelog I see a few things that *might* be it. including:

c5d3825 (#11740) Wait on the handle from the PROCESS_INFORMATION structure
4f493a3 (#12564) Stub CHILDSTATUS in all test cases, not just failure cases

Either way, lets call this Confirmed for now, but I think it may prove to be fixed already in quantal.

Changed in puppet (Ubuntu):
status: New → Confirmed
Revision history for this message
Clint Byrum (clint-fewbar) wrote :

Have run into this quite a few times now on precise. Its really quite dangerous as puppet just randomly seems to fail.

I've looked through the changelog of puppet 2.7.12 and can't find any obvious reason this doesn't happen, so the next step is likely to find a reliable reproducer (tough, probably a race) and then bisect it.

Changed in puppet (Ubuntu):
importance: Medium → High
Revision history for this message
Adrien (adrienburie) wrote :

Thanks James & Clint for your answers.

I would like to add some details:

Currently, as a temporary solution, we use crontab to kill the process at 00h00 and start again at 00h01. With that, we are sure that the agent works every days.

With my co-worker, we have updated a server (Ubuntu 12.04 LTS) which had the stable version (2.7.11) with the version of quantal (2.7.18, we used pin preferences for that), and we have the same problem.. Nothing wrong in syslog

For example, saturday, the last execution of the puppet was at 07h00 AM, and the next at 00h00 sunday (thanks to crontab)..

I hope this information can help you..!

Regards

Adrien

Revision history for this message
Michael Stahnke (stahnma) wrote :

The upstream bug should actually be http://projects.puppetlabs.com/issues/10418. The originally referenced one is a duplicate.

Revision history for this message
Haw Loeung (hloeung) wrote :

Might also be bug #995719. Can you try remove process_name.rb ?

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.