Let me also add that I have several servers that have the same exact
role in their node definitions and don't have any problem with
updates.
On Nov 10, 11:40 am, PBWebGuy pbweb...@gmail.com wrote:
I have one of 18+ servers in an environment that just started having a
problem when attempting
A few more findings:
1. I removed the /var/lib/puppet/clientbucket and ran update manually
it worked one time. Repeating this does not work.
2. Here is the Trace of the puppetd command:
/usr/lib/ruby/1.8/net/http.rb:586:in `connect'
/usr/lib/ruby/1.8/net/http.rb:586:in `connect'
Make sure that you get output from the puppetmaster for some things. It's
easy to look at the wrong log if you are using passenger.
I've seen a similar error problem when the puppetmaster is overloaded so the
client times out. For me, this is when It gets all 50 clients ask for a
I have continued troubleshooting this and have gone as far as:
1. Removed all code from the nodes.pp for this particular node.
2. Deleted the /var/lib/puppet directory on the node and resigned with
the PM. Note that the signing failed repeatedly as well where the PM
did not report any messages
After 6 hours of troubleshooting, we found that there was a process
(Alfresco/Tomcat) running on the node that was consuming large amounts
of the CPU. Running 'top', there was a sustained load of 2.x. As
soon as I killed the process, Puppet started running perfectly! This
node is a VM in a