Mark Miller created SOLR-8121: --------------------------------- Summary: It looks like ChaosMonkeySafeLeader test can fail with replica inconsistency because waitForThingsToLevelOut can pass while state is still changing. Key: SOLR-8121 URL: https://issues.apache.org/jira/browse/SOLR-8121 Project: Solr Issue Type: Test Reporter: Mark Miller Priority: Minor
If this passes and things have not actually leveled out yet, we do a strict check that fails without waiting as we wanted for a stable cluster. My quick fix has been to call waitForThingsToLevelOut, wait a couple seconds, then call it again. It's rare enough to hit this that I think that may bring the false cases below what we can see. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org