Hi solr devs/users, I'm setting up solr slave test server now (duo core with 2GB ram), and I'm running into some issues with replication. The index being replication is in the size of about 10G now, and the snappuller and snapinstaller are scheduled to run every 30 mins using crontab.
Upon running snapinstaller, cache would autowarm itself and I would run into OOM pretty quickly. For now, I have turned off autowarm, new searcher listener, and first searcher listener JUST to get the new searcher registered upon new index being installed. I'm able to get this part going so far. I also have some test script to query against the slave server; however, whenever during snapinstall, OOM would occur and the server is not very responsive (even with autowarm disabled). After a while (like couple minutes), the server can respond again. Is this expected? What can I do to make it not run into OOM during snapinstall with autowarm enabled? Does anyone know how the cache in solr/lucene utilize the heap in terms of generation (young vs tenured)? I suspect that setting the right ratio between the young and tenured generation could keep the responsiveness, especially during GC. I have set the heap size to 1.5GB out of the 2GB physical ram. Any help is appreciated. Thanks! -Gaku -- View this message in context: http://www.nabble.com/solr-slave-configuration-help-tp17583642p17583642.html Sent from the Solr - User mailing list archive at Nabble.com.
