How do you get a manifest to work when it stops after 300 seconds even when the timeout is disabled?

asked 2017-01-10 18:17:00 -0500

Conventional gravatar image

Why would a Puppet Manifest stop running after 300 seconds with no error?

I have Puppet Master 4.x (enterprise edition) on RHEL 6.7. I have Puppet Agent on RHEL 6.7. I have a manifest that runs and appears to work. It says "Catalog applied" after approximately 300 seconds. The problem is that it isn't working but there are no logs or messages indicating that. The manifest transfers a bash script and runs it. The bash script would otherwise take 10 minutes to run.

I changed the timeout of the manifest to 3600 and 0. With three different runs of this manifest (with no timeout stanza, or setting the timeout to 3600 or 0), the results are the same. The Puppet Master logs (in /var/log/...) do not give me an idea of what is wrong. I do not see the "timeout" error even with "puppet agent -t -d" on the Puppet Agent server. The run of the Puppet agent providing a false positive that the transaction completed and the catalog applied.

Why would a Puppet Manifest quit running after 300 seconds and not provide an error message?

edit retag flag offensive close merge delete

Comments

1

Could you provide the source code. I'm not aware that you can timeout a manifest (a *.pp file). Why are you running a sh-script that takes ten minutes? Does _it_ _configure_ the system? Or do something else non-configuration related? Can you invoke your script on the command line w/o problems?

Kai Burghardt gravatar imageKai Burghardt ( 2017-01-10 19:38:20 -0500 )edit