On 12/03/2011 04:30 PM, Jeff Sussna wrote:
I am considering installing a Puppet agent on a legacy server running an old, crappy, slow, leaky, yet important Java app. Nervous about introducing anything that might (further) destabilize the app. What are people's experiences with Puppet in terms of resource usage (CPU/ memory/network/etc)? For starters will likely run it on the standard daemon schedule. Ultimately may want to control it with MCollective though.
Puppet isn't the fastest or lightest service around. It is, after all, an interpreted language implemented in an interpreted language. I run it on a handful of small VMs. Resource usage seems to be mostly a function of the number of resources. The worst case I have is my Nagios monitoring server, which has about 500 resources and 256 MB of RAM. It runs Apache, Nagios, and Puppet agent. When the agent runs, the RAM is pretty much exhausted, and the run takes a good chunk of the 1.2 GHz CPU for about one minute. Would probably be faster if it wasn't swapping.
However, if your hardware is a little less minimal, or you have fewer resources, it's likely you won't notice. I have seen the puppet agent daemon consume an abnormal amount of RAM as if it's leaking from time to time. That could probably be avoided by running it from another scheduling tool, or running something more recent than the 2.6.2 that's in Debian Stable.
Most of my other nodes are similar, but have far fewer resources, so I don't even notice the agent running. The master runs with similar resources and can keep up with my 20 or so agents easily.
-- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
