Issue
This is a known issue related to Solr: https://issues.apache.org/jira/browse/SOLR-14431
that affects at least down to 8.1.1 and is fixed in 8.6 that makes it so replicas can get out of sync with each other.
Environment
Solr 8.1.1 -> 8.5+
Fixed: 8.6
Cause
This happens as operations "stall" due to the threads attempting to execute them blocking forever waiting for a ReentrantReadWriteLock in DefaultSolrCoreState that will never be released.
Further reading is here: https://issues.apache.org/jira/browse/SOLR-14431
Resolution
There is a workaround without having to upgrade. You can add the following configuration into the solrconfig.xml. By explicitly declaring the /admin/segments handler in solrconfig.xml with an invariants param that requests additional info, forcing it down a code path where it uses the IndexWriter, and decrements the ref-count, releasing the lock.
<requestHandler name="/admin/segments" class="solr.SegmentsInfoRequestHandler">
<!-- work around for https://issues.apache.org/jira/browse/SOLR-14431 -->
<lst name="invariants">
<bool name="coreInfo">true</bool>
</lst>
</requestHandler>
Comments
0 comments
Article is closed for comments.