On 4/7/2021 12:29 AM, jay harkhani wrote:
Query : https://drive.google.com/file/d/1RHnIVsK7m-T8_3it97jLquyfBFm3OGfj/view
Schema File:
https://drive.google.com/file/d/1SoKpucgtws3teW0QbV7UKg9xiV1cYuXJ/view
For better performance we used term query but still not getting quick result.
To analyze further more we found that if we remove grouping field or we reduce
no. of ids from query, it executes fast. Is there any better option for fast
execution when large query or more ids pass in one field?
Can you provide a screenshot like the ones mentioned here?
https://cwiki.apache.org/confluence/display/SOLR/SolrPerformanceProblems#SolrPerformanceProblems-Askingforhelponamemory/performanceissue
Have you fiddled with the GC tuning, or is it using the default 6.x CMS
settings? Can you provide the full commandline for the Solr process
once it gets going?
What Java vendor and version do you have?
If you can provide all the GC logs that have been saved by Solr, that
could be helpful.
Querying that many IDs is going to be slow. This is why the
MaxBoelanClauses value defaults to 1024. The term query is faster than
the standard boolean query.
Do you know the cardinality of the summary field, and specifically the
cardinality of docs that match the following part of your query:
(summary:((test*) )
Wildcard queries like this are another reason for things going slowly.
Thanks,
Shawn
