I intentionally broke my puppet_node_classifier to see exit codes.

See also: http://projects.reductivelabs.com/issues/2162

puppetd prints out that there is a problem:

/usr/sbin/puppetd —server=01.demo.test —test

err: Could not retrieve catalog: Could not find node ‘01.demo.test’; cannot compile warning: Not using cache on failed catalog

puppetrun, however, merrily thinks that everything is okay:

puppetrun —host=01.demo.test —foreground

Failed to load ruby LDAP library. LDAP functionality will not be available Triggering 01.demo.test 01.demo.test finished with exit code 0 Finished

The exit code I’ve submitted in the previous bug, but I think puppetrun should give me an indication of whether there were any problems with the run.

Context, in case it’s helpful:

I’m controlling puppet via a web interface, and want to take an action across a cluster of machines, like “upgrade these servers to a new version of an application”. puppet_node_classifier pulls from a database, and then the nodes update themselves according to the rules laid out. I want to be able to indicate in the web interface whether the upgrade was successful or not. Currently, it looks to me that puppetrun is only reporting whether or not the trigger of a manual run was successful, and not whether the manual run itself was successful.

Pete