[ https://issues.apache.org/jira/browse/CASSANDRA-15402?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17626500#comment-17626500 ]
Stefan Miklosovic commented on CASSANDRA-15402: ----------------------------------------------- Thanks for the incorporation of my feedback, [~maxwellguo]. One last problem I have is that the logic should be inverted. I think we need to update cassandra.yaml as well. There is this currently: {code} # Set to true to have Cassandra create a hard link to each sstable # flushed or streamed locally in a backups/ subdirectory of the # keyspace data. Removing these links is the operator's # responsibility. incremental_backups: false {code} It means that we are changing the behavior, because if you set incremental_backups to true, it will not start to backup incrementally as it was before, because you have set Option.INCREMENTAL_BACKUPS to be false by default. Hence, I think that in order to keep same logic as it was, the default value for table should be true instead of false. That means that you opt-out if you want but you would be opted-in by default. We need to mention in cassandra.yaml that one can optionally turn off incremental backups by setting table parameter to false if incremental_backups is set to true. > Make incremental backup configurable per keyspace and table > ----------------------------------------------------------- > > Key: CASSANDRA-15402 > URL: https://issues.apache.org/jira/browse/CASSANDRA-15402 > Project: Cassandra > Issue Type: Improvement > Components: Local/Other, Tool/nodetool > Reporter: maxwellguo > Assignee: maxwellguo > Priority: Normal > Labels: pull-request-available > Fix For: 4.x > > Time Spent: 0.5h > Remaining Estimate: 0h > > We know that when we do backup for cassandra, we can do full bakcup with > snapshot, when we need to backup incremental data , incremental_backup can do > some help . > For snapshot we can just make snapshot for some keyspace/table. but > incremental backup will do all data backup(make hard link for all sstable > that flush from memtable or streaming ). we also know that commitlog's replay > can do some keyspace /table 's filter. > So I think for incremental backup we also need some filter for it.After all > not all keyspace/table is so important to do backup. -- This message was sent by Atlassian Jira (v8.20.10#820010) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org