[ https://issues.apache.org/jira/browse/CASSANDRA-14332?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16436783#comment-16436783 ]
Kurt Greaves commented on CASSANDRA-14332: ------------------------------------------ |[3.0|https://github.com/apache/cassandra/compare/cassandra-3.0...kgreav:13797-r-3.0]|[3.11|https://github.com/apache/cassandra/compare/cassandra-3.11...kgreav:13797-r-3.11]| |[utests|https://circleci.com/gh/kgreav/cassandra/152]|[utests|https://circleci.com/gh/kgreav/cassandra/154]| I think reverting this is the correct solution for 3.0 and 3.x. Changing defaults in minors while in this case is probably fine, it's not good practice. Backporting CASSANDRA-13521 completely is also a bit questionable. The original problem from CASSANDRA-13797 of the interrupted exceptions is only very minor, and likely not possible outside of CCM clusters. > Fix unbounded validation compactions on repair > ---------------------------------------------- > > Key: CASSANDRA-14332 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14332 > Project: Cassandra > Issue Type: Bug > Reporter: Kurt Greaves > Priority: Major > > After CASSANDRA-13797 it's possible to cause unbounded, simultaneous > validation compactions as we no longer wait for validations to finish. > Potential fix is to have a sane default for the # of concurrent validation > compactions by backporting CASSANDRA-13521 and setting a sane default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org