pluginsync not working with 3.7.0 on RHEL6

[Adam Winberg]

After upgrading clients to 3.7.0 on RHEL6 my puppet runs error out with:

------------------

Info: Retrieving pluginfacts

Info: Retrieving plugin

Info: Loading facts

Error: Could not retrieve catalog from remote server: end of file reached

Warning: Not using cache on failed catalog

Error: Could not retrieve catalog; skipping run

------------------

It works fine with 3.6.2 on the clients and 3.7.0 on the masters, the error occurs when running clients with 3.7.0. I can work around it by settins pluginsync to false, so there seems to be a problem with this option. On my RHEL7 systems it works fine (same config and same puppet version).

Anyone have an idea of whats going one?

//Adam

[Adam Winberg]

think i found the problem, I recently installed the stdlib module and knew that it would mean that I had duplicate functions since i had already installed the dirname.rb function in another module. However when i tested it (on 3.6.2) it worked fine, puppet chose one of the functions and did not complain about it. However, 3.7.0 is not as forgiving and crashes without saying a word about why. I will file a bug to get a better error message for this. 

//Adam

--
You received this message because you are subscribed to the Google Groups "Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to puppet-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/38ae1605-8106-4cdb-8c34-74f5fd43b503%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Adam Winberg]

I spoke to soon... Removing, or rather making a change in the plugins, made the puppetrun work once, but the followin run fails with the same error. So I am still very open to ideas. 

//Adam

[Adam Winberg]

Found the problem for real, i hope. puppet 3.7 has a default setting of 4s for "http_keepalive_timeout". This setting should not be higher than the Apache 'KeepAliveTimeout' setting on the puppetmaster (running passenger on apache). In my apache, we have for some reason set a really low KeepAliveTimeout setting and by adjusting this to '5' as to get it higher than puppet default the problem went away. So, as always, the problem arise due to some specific configuration in my environment. over and out. 

//Adam