magibney opened a new pull request, #1986: URL: https://github.com/apache/solr/pull/1986
https://issues.apache.org/jira/browse/SOLR-17015 The extant test for AbstractMoveReplicaTestBase didn't cover PRS at all, and was also a bit naive in the sense that it tested replica moves on a static system, with no concurrent updates coming in. This PR enhances the main test to concurrently index and commit to the collection as the MoveReplicas are happening, pausing enough to verify that no docs have been dropped in the process. It also randomly enables/disabled PRS for the test run. The more robust testing uncovered an optimization in ActiveReplicaWatcher (to work around behavior that's fixed by #1964) that assumes non-PRS-enabled collections. This optimization ended up causing PRS moves to fail in some circumstances (reliably triggered by the enhanced test). This PR removes that optimization, which after merging #1964 should be irrelevant anyway. Unsurprisingly, the more intensive concurrent indexing and query load increases the latency of the test, in the worst case ~2x, but often actually slightly _faster_ for PRS, and on average ~20% slower (based on beasting the test). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org For additional commands, e-mail: issues-h...@solr.apache.org
