[
https://issues.apache.org/jira/browse/SOLR-16758?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17718568#comment-17718568
]
Kira Traynor commented on SOLR-16758:
-------------------------------------
Yes that is the query that I have been running. It was chosen just because it
has the longest latency so that it would be the easiest to compare. The way
that it was discovered was with different queries that contained
traversalFilters and were not just * : * queries.
I also want to include the test that I was running with the local instances of
solr using the Http2SolrClient because with the GraphQueryTest that I included
I also was not able to get any distinct differences in the two performances.
When running the test I got a lot of variability ranging from 200ms to over
700ms for the average query latency which is why I wanted to run it on a docker
container.
Also, thanks for pointing out the optimization. The indices that the latency
differences were initially found on were optimized down to one segment. The
testing I have done with the generated index has not been optimized and the
testing has been done with 9 segments. I've done some more testing to compare
with different amounts of segments and deleted docs.
Optimized: 228ms patch vs 238.0ms baseline
Not Optimized (9.09% deleted docs and 9 segments): 271ms patch vs 336.0ms
baseline
Not Optimized (9.09% deleted docs and 5 segments): 303ms patch vs. 328ms
baseline
Not Optimized (25.93% and 12 segments): 274ms patch vs. 334ms baseline
[^QueryLatencySolrTest.java]
> GraphQuery Performing Slower in Solr 9
> --------------------------------------
>
> Key: SOLR-16758
> URL: https://issues.apache.org/jira/browse/SOLR-16758
> Project: Solr
> Issue Type: Bug
> Reporter: Kira Traynor
> Priority: Major
> Attachments: GraphQuery.patch, QueryLatencySolrTest.java
>
>
> We found that the changes made in SOLR-14185 with the GraphQuery slowed down
> the performance of the graph join by around 50%. Reverting the changes made
> in that pr to use a `Filter` object to get the `DocIdSet` rather than using a
> `DocIdSetIterator`, sped the query times up in our tests to be more similar
> to what we were seeing with the Solr 8 graph joins.
> The same change was made to the JoinQuery class so possibly slower
> performance can be found there as well.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
