Hello Jan, On Tue, May 7, 2024, 12:05 PM Jan Høydahl <jan....@cominvent.com> wrote:
> Hi, > > If you have not tuned your heap, it will default to 512m. THe lines in > solr.in.sh prefixed with hash # are comments and do nothing. > If you experience OOME during use that is likely not enough for your use > case. > So edit the file, remove the hash and set SOLR_HEAP=1024m to give yourself > some more headroom, then restart the service. > I will try this but do bear in mind that I have 4Gb to play with. The server is also running other workloads (database, webserver etc). Just fyi, on the weekend I experimentally set both of these to be 256m SOLR_HEAP SOLR_JAVA_MEM The OOM killer issue came back. You may want to add some monitoring < > https://solr.apache.org/guide/solr/latest/deployment-guide/monitoring-with-prometheus-and-grafana.html> > on your service to gain insight into hotspots and performance to do a more > informed tuning < > https://solr.apache.org/guide/solr/latest/deployment-guide/jvm-settings.html#choosing-memory-heap-settings > >. > Since 9.2 the solr process should actually exit on OOM, thus if you > installed Solr as a daemon service it should restart automatically? > > Jan > I'm running Solr 9.4.x. I installed it via the official tarball and some parts of the installation were automated. It's managed via systemd. There does seem to be a way to auto restart managed services on termination but I want to try "one thing at a time" and see how much I can improve the OOM issue first.
