On Fri, Dec 14, 2012 at 11:00 AM, scukao...@gmail.com <scukao...@gmail.com>wrote:
> hi, all > > I have a master for about 3000 clients, runinternal is 20 min. > A problem is that puppetdb's mq is too long so that data in postgre is the > OLD version. any suggests to improve performance ? > We'll need some more information in order to get a complete picture of what's going on. Can you send a screenshot of your puppetdb web console after you've had it up for a few minutes? In particular, we'd need to see the rate at which your queue is growing, metrics around command processing time and enqueue time, as well as information about catalog & resource duplication rates. Also, what version of Postgres are you using, and have you tuned any of the settings? Lastly, is the puppetdb system experiencing a lot of iowait, or is it otherwised blocked on i/o or system time? With that many nodes and a 20 minute runinterval, you'll be sending ~2.5 catalogs every second to puppetdb...it's not insurmountable, but we'll need to look at the whole system in totality to see where the bottleneck is. :) The first-order thing I'd look at is your catalog duplication rate...if that's low, then puppetdb is doing a lot of reads/writes to the database for every catalog received. Improving that number will help let puppetdb short-circuit writing to disk at all, which should improve throughput considerably. You can also come and find us on IRC at #puppet on Freenode; that may make for a faster back-and-forth. Thanks! -- deepak / Puppet Labs -- 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.
