Hi David, Re: your comment on SOLR-17018 <https://issues.apache.org/jira/browse/SOLR-17018> about the latest implementation to address query timeout in 9.3.0 SOLR-16693 <https://issues.apache.org/jira/browse/SOLR-16693>.
Can you please clarify the use of query timeout vs timeAllowed again? I read the note <https://cwiki.apache.org/confluence/display/SOLR/ReleaseNote9_3_0> but it is not clear if SOLR-16693 <https://issues.apache.org/jira/browse/SOLR-16693> improvement helps in reducing the latency of standard queries that are limited to searching with "boost" and/or "bf" and *without* additional search components? Without "timeAllowed" param, If the query includes an expensive facet or expensive reranker(ltr) query in this scenario, will it continue processing until its completion? If yes then this improvement does not resolve the issue reported on SOLR-17018 <https://issues.apache.org/jira/browse/SOLR-17018>. The facets or reranker that are processed on top of the main search results, if they cannot be processed further, within time allowed, then the better solution is what "timeAllowed" currently implements, to stop the processing and return partial results. Would the future improvement be around timeAllowed making use of the implementation of TimeLimitingBulkScorer instead of ExitableDirectoryReader and LTR component implementation timeAllowed SOLR-14607 <https://issues.apache.org/jira/browse/SOLR-14607>? Thank you all for looking into this one. Rajani On Wed, Oct 11, 2023 at 2:47 PM rajani m <rajinima...@gmail.com> wrote: > Hi Shawn, > > Thank you for looking into this. I have updated the jira with further > investigation https://issues.apache.org/jira/browse/SOLR-17018. Please > take a look and let me know what you all think. > > On Mon, Oct 9, 2023 at 2:22 PM Shawn Heisey <apa...@elyograg.org.invalid> > wrote: > >> On 10/8/23 18:24, rajani m wrote: >> > I have filed a Jira SOLR-17018 >> > <https://issues.apache.org/jira/browse/SOLR-17018> explaining the >> issue and >> > think it is just a certain type of queries causing the issue and not any >> > bug, but still looking into it to see if anything else. >> >> You need to find out what resource was actually exhausted. It is not >> always memory. Solr 9.1.1 may not always log the reason for the OOME, >> and the gist you shared does not include the OutOfMemoryError exception. >> >> Solr 9.2.0 and later will always log the reason in a crash log. Before >> doing ANYTHING, you must know which resource was exhausted, or you could >> be chasing the wrong problem. >> >> Once you determine which resource was exhausted, there are exactly two >> ways to deal with it: >> >> * Increase the available amount of the exhausted resource. >> * Change the config or the requests you are sending so Solr uses less of >> that resource. >> >> The second option is not always possible, and whether you can do the >> first option depends on the hardware. >> >> Thanks, >> Shawn >> >>
