Jon Haddad created CASSANDRA-19589:
--------------------------------------
Summary: Improve migration to incremental repair
Key: CASSANDRA-19589
URL: https://issues.apache.org/jira/browse/CASSANDRA-19589
Project: Cassandra
Issue Type: New Feature
Reporter: Jon Haddad
In order to make it easier to migrate from subrange repair to incremental
repair, I think we should support an option to repair data already marked as
repaired.
The process would require the user to do the following:
1. A new command to mark all data for a table as repaired, cluster wide.
2. Going forward, the user runs incremental repairs on a regular basis
3. To ensure the data that was written prior to step 1 is consistent, the user
would now run subrange repair on *only* the data marked as repaired.
This would allow the repaired data to remain marked as repaired while also
ensuring it's consistent. It also makes it easy for the user to have a regular
repair job on the already repaired data to ensure that things are consistent,
or after replacing a node from a backup that may have a different set of
repaired data from the rest of the cluster.
--
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
