Sorry Yonik, I hope this didn't come off as criticism. Far from it. We are very happy with the performance we are getting. I just happen to be the performance junkie trying to get every little bit out.
That being said, I'm happy to hear it's going to get even better! -Todd -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Yonik Seeley Sent: Tuesday, October 14, 2008 1:38 PM To: solr-user@lucene.apache.org Subject: Re: Practical number of Solr instances per machine On Tue, Oct 14, 2008 at 4:29 PM, Feak, Todd <[EMAIL PROTECTED]> wrote: > In our load testing, the limit for utilizing all of the processor time > on a box was locking (synchronize, mutex, monitor, pick one). There were > a couple of locking points that we saw. > > 1. Lucene's locking on the index for simultaneous read/write protection. > 2. Solr's locking on the LRUCaches for update protection. Luckily, both of these are very close to being improved: 1. Lucene 2.4 has NIO support (lockless) except for Windows, and there is already a Solr patch to add support for that. 2. Solr already has a patch (soon to be committed) for an LRUCache based on ConcurrentHashMap that should work better with multiple CPUs. -Yonik
