Mysterious puppet agent hangs, stale puppetdlock

Recently, there has been some strangeness happening with the puppet agent process (puppetd), where during a run, the agent gets stuck and does not finish removing the lock file. This results in a useless puppet agent, until it is restarted. This can become a big problem for many reasons, mainly that nobody wants to have to repeatedly log in to every machine to restart the puppetd service.

This issue is discussed at length in this thread

There is more than one way to solve this problem, so below I'll detail my solution. Keep in mind that this is not intended to be a permanent fix, but merely a way of keeping the machines running until a more elegant solution is implemented.

My solution is a simple cron/script type. Every 5 minutes, I run a script to check the last modified time of the puppetdlock file, and compare it to the current time. If it is older than 5 minutes, I call a puppetd restart. That's it!

# File Name:
# Author:        Ryan Uber <>
# Description:   This is a *TEMPORARY* fix to the common puppetdlock issue.
#                Running this script checks if the lock file is more than 5
#                minutes old, and if it is, restarts the puppet process.
# See Also:

# Path to the puppetdlock file

# Number of minutes since last modification that indicate a stale lock file

# UNIX timestamp of last puppetdlock modification
MODIFIED=$(stat -c %Y ${FILE})

# Current UNIX timestamp
NOW=$(date +%s)

# Check if file is stale or not and restart if it is.
if [ $((${NOW}-${MODIFIED})) -gt $((${STALEMIN}*60)) ]
    /sbin/service puppet restart


See one of my other posts if you are interested in how I went about pushing out the cron job.