About week ago, we moved our production puppet server to an older IBM
OpenPower Server with 8 cores and 8GB ram. We use Ubuntu 8.04 LTS for
the PowerPC (the most recent version for this hardware). We have
puppet 0.24.8, Apache+ssl, Passenger 2.2.2 and Rack 1.0.
We manage 780 identically configured compute nodes and gateways with a
mix of ppc and x86 cpu''s running either CentOS 5.3, or RHEL 5.
On the server, all the ruby software including puppet and facter,
along with ruby itself, was built from source. Facter uniformly
misreports the number of processors on the PowerPC boxes, and a group
of dual processor x86. We are testing two facter additions to correct
the problem, since the ppc''s have a few small but important
differences from the x86''s, namely the way nrpe is used.
We used the defaults provided by a stock install including having all
780 boxes check-in every 30 minutes.
We now have nearly 30 modules managing everything from autofs to yum.
We use torque and moab for scheduling user jobs, so we use puppet to
maintain the config files, directories and services needed for that as
well.
On the server we typically see:
20:46:13 up 6 days, 10:14, 1 user, load average: 0.00, 0.63, 0.89
Finally, Passenger is configured:
PassengerPoolIdleTime 300
PassengerUseGlobalQueue on
PassengerHighPerformance on
PassengerMaxPoolSize 15
Cheers--
Charles
Advanced Computing Center for Research and Education
Vanderbilt University
--~--~---------~--~----~------------~-------~--~----~
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+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---