[ https://issues.apache.org/jira/browse/CASSANDRA-15454?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16997483#comment-16997483 ]
Aleksey Yeschenko commented on CASSANDRA-15454: ----------------------------------------------- Cheers, committed to 4.0 as [8202845facd741f01fbfbbec93d6a3c8e6078644|https://github.com/apache/cassandra/commit/8202845facd741f01fbfbbec93d6a3c8e6078644] > Update generations and update comments for altered distributed system > keyspaces > ------------------------------------------------------------------------------- > > Key: CASSANDRA-15454 > URL: https://issues.apache.org/jira/browse/CASSANDRA-15454 > Project: Cassandra > Issue Type: Bug > Components: Cluster/Schema > Reporter: Aleksey Yeschenko > Assignee: Aleksey Yeschenko > Priority: Normal > Fix For: 4.0 > > > In 4.0, we set {{memtable_flush_period_in_ms}} to 0 (from 1 hour previously), > and changed default chunk length from 64KiB to 16KiB. On startup we would try > to override existing schema definitions for {{system_traces}}, > {{system_distributed}}, and {{system_auth}} to reflect these changes, but > because these are changes to smaller values, while timestamps remain the > same, these will never override existing, pre-upgrade values from 3.0 or > 3.11. As a result, a schema migration request will be pushed by a node on > every bounce to every replica, causing unnecessary and ultimately fruitless > work. > Bumping the generations fixes the issue, and documenting the changes doesn't > hurt either. -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org