python process hogged the cpu in 12.04

Bug #1003138 reported by Adrien Lamothe
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
juju (Ubuntu)
New
High
Unassigned

Bug Description

System was unresponsive. A 'top' command revealed a process named 'python' using 99% cpu. I believe todays 12.04 updates were python programs for single sign on, so perhaps that had something to do with it. I've rebooted and so far the system is OK again.

Aditya V (kroq-gar78)
affects: ubuntu → python-defaults (Ubuntu)
Revision history for this message
Julian Taylor (jtaylor) wrote :

if you can reproduce the issue please provide the full command line of the spinning python process:

cat /proc/<pid>/cmdline | tr '\0' '\n'

where <pid> is the process id shown in the first column of top

Changed in python-defaults (Ubuntu):
status: New → Incomplete
Revision history for this message
Adrien Lamothe (alamo) wrote :

It happened again after today's updates (which included some python stuff.) I didn't run that command, too busy with other things to remember that. You may want to prompt users for reboot after installing those types of python packages.

Revision history for this message
Adrien Lamothe (alamo) wrote :

Just installed 12.04 on a different computer and then updated it, in an effort to recreate the problem. Unfortunately the problem didn't occur. Will just have to wait until it happens again on my primary machine.

Revision history for this message
Adrien Lamothe (alamo) wrote :

OK, it happened again. Looks like it is JuJu causing the problem. Is it a coincidence it happened after updates?

Output of cat /proc/<pid>/cmdline | tr '\0' '\n':

-m
juju.agents.machine
--nodaemon
--logfile
/home/adrien/juju/adrien-sample/machine-agent.log
--session-file
/var/run/juju/adrien-sample-machine-agent.zksession

Julian Taylor (jtaylor)
affects: python-defaults (Ubuntu) → juju (Ubuntu)
Changed in juju (Ubuntu):
status: Incomplete → New
Changed in juju (Ubuntu):
importance: Undecided → High
Revision history for this message
Adrien Lamothe (alamo) wrote :

Just happened again, have attached a screen shot of the 'top' output. Also:

~$ cat /proc/1007/cmdline | tr '\0' '\n'
/usr/bin/python
-m
juju.agents.machine
--nodaemon
--logfile
/home/alamo/juju/alamo-sample/machine-agent.log
--session-file
/var/run/juju/alamo-sample-machine-agent.zksession

Revision history for this message
Adrien Lamothe (alamo) wrote :

It appears the bug also output huge amounts of data to the machine-agent.log file, ~ 93 GB, which completely ate up my free disk space. This happened gradually over a period of time. I've attached a screen shot showing this.

It is important to note that during the period where Juju was running amok, I never once used it. It was merely installed on my computer.

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

Hi Adrien, this is a duplicate of bug #1006553 , which will be fixed by not starting the agents on reboot (since the lxc containers and zookeeper are not also resurrected).

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.