On 8/6/2017 10:29 PM, hawk....@139.com wrote: > We found the problem is caused by the delete command. The request is used to > delete document by id. > > url --> http://10.91.1.120:8900/solr/taoke/update?&commit=true&wt=json > body --> {"delete":["20ec36ade0ca4da3bcd78269e2300f6f"]} > > When we send over 3000 requests, the Solr starts to give OOM exceptions.
Do you have the full exception with stacktrace for those OOM? I'm curious exactly what resource ran out, whether it was heap or something else. When java programs throw OOME, the program's execution is usually completely unpredictable from that point on, because something that the program tried to do did not happen. Whatever the program tries to do next probably depends on the action that failed. This unpredictability is why Solr on non-windows systems will self-terminate when OOME is encountered. It's the only safe action to take. There is an issue to bring the same self-termination on OOME to Solr running on Windows. If the OOME was due to heap space, there are exactly two ways to deal with that. You can find info about it here: https://wiki.apache.org/solr/SolrPerformanceProblems#Java_Heap Thanks, Shawn