Yeah, lots of evictions and tiny caches. Why not increase them? It looks like you have memory to spare. And since you reopen the searcher so often, you can play with increasing the warm-up time if you want to preserve more cached items from the previous searcher.
Evictions are measured in the number of occurrences, not bytes. Otis -- Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch ----- Original Message ---- From: Jonathan Ariel <[EMAIL PROTECTED]> To: solr-user@lucene.apache.org Sent: Tuesday, April 15, 2008 11:03:53 AM Subject: Re: too many queries? Thanks. It should be around lookups*1.5, right? Is this measured in bytes? On Tue, Apr 15, 2008 at 11:26 AM, Erik Hatcher <[EMAIL PROTECTED]> wrote: > Filter cache evictions are a big red flag. Try bumping up the size of > your filter cache to avoid regenerating filters. > > Erik > > > > > On Apr 15, 2008, at 8:38 AM, Jonathan Ariel wrote: > > > filterCache > > autowarmCount=256 > > lookups : 24241 > > hits : 21575 > > hitratio : 0.89 > > inserts : 3708 > > evictions : 3155 > > size : 512 > > cumulative_lookups : 2662056 > > cumulative_hits : 2355474 > > cumulative_hitratio : 0.88 > > cumulative_inserts : 382039 > > cumulative_evictions : 365038 > > > > The CPU usage is usually 50%. > > I give the JVM "java -server -Xmx2048m" when I start Solr. > > > > Thanks! > > > > Jonathan > > > > > > > > > > On Mon, Apr 14, 2008 at 8:24 PM, Otis Gospodnetic < > > [EMAIL PROTECTED]> wrote: > > > > It's hard to tell from the info given, though something doesn't sound > > > ideal. Even if Solr's caching doesn't help, with only 4M documents, > > > your > > > Solr search slaves should be able to keep the whole index in RAM, > > > assuming > > > your index is not huge. > > > > > > How large is the index? (GB on disk) > > > Is it optimized? > > > How often is it changed on the master - i.e. how often does your > > > Searcher > > > need to be reopened? > > > What are cache hits and evictions like (Solr admin page)? > > > What are cache sizes like and how is the warm-up configured? > > > Is there any IO on the slaves? (run vmstat or iostat or some such) > > > How is the CPU usage looking? > > > > > > Otis > > > -- > > > Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch > > > > > > ----- Original Message ---- > > > From: Jonathan Ariel <[EMAIL PROTECTED]> > > > To: solr-user@lucene.apache.org > > > Sent: Monday, April 14, 2008 5:50:08 PM > > > Subject: too many queries? > > > > > > Hi, > > > I have some questions about performance for you guys. > > > So basically I have 2 slave solr servers and 1 master solr server load > > > balanced and around 100 request/second, aprox. 50 request per second > > > per > > > solr server. > > > My index is about 4 million documents and the average query response > > > time > > > is > > > 0.6 seconds, retrieving just 4 documents per query. > > > What happens is that there are too many request to Solr and every > > > second > > > is > > > getting bigger, so eventually my site stops working. > > > > > > I don't know if this stats are enough to tell if the servers are > > > supposed > > > to > > > handle this amount of request. Maybe it's a configuration problem. I > > > don't > > > think that caching in solr would help in this case because all the > > > queries > > > are different (I'm not sure how caching works but if it's per query it > > > won't > > > help much in this case). > > > > > > Any thoughts about this? > > > > > > Thanks! > > > > > > Jonathan > > > > > > > > > > > > > > > >
