dcapwell commented on code in PR #3743: URL: https://github.com/apache/cassandra/pull/3743#discussion_r1917517216
########## src/java/org/apache/cassandra/index/accord/ParticipantsJournalIndex.java: ########## Review Comment: > In this class, in #getInitializationTask, there is forceBlockingFlush task. Journal tables do not have memtables, so I would either remove this code, or replace it with the logic that would assert that memtables are empty. So, memtable and 2i requires these stuff, so when i started asserting that memtables are never touched some tests started to fail... stuff like `TRUNCATE` break... i rather not try to block at this layer as it causes issues. > What would generally happen if the index is non-queriable? We probably need to throw, since if there's a chance we can return empty resultset, it can be a severe correctness violation. on startup this won't be an issue, this would only be an issue if the user does `nodetool rebuildindex system_accord.journal`. If the index isn't up we kinda just broke the clusters ability to function... For now ill add a check that the index is up -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
