[
https://issues.apache.org/jira/browse/CASSANDRA-5351?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13774348#comment-13774348
]
Marcus Eriksson commented on CASSANDRA-5351:
--------------------------------------------
Imo, keeping a separate set of levels for unrepaired data would be the way to
go (or running 2 parallel STCS if thats the CS you run). I guess this would
make reads a bit slower, but imo it is worth it, reads are _alot_ slower
during/after repairs today anyway, amortizing that cost would make things more
predictable.
With LCS you would have a bunch of nicely leveled newly flushed data, which is
not yet repaired. Then, when repair is run, we would drop that data + any
streamed data to level 0 in the repaired-data-LCS. And since the data is
already partitioned nicely we would get very parallel L0 -> L1 compactions in
the 'repaired'-LCS.
also, i guess this would make 'repair -pr' not work since we would need to
repair all data on the node (which wouldn't be a problem since the amount of
data to repair is tiny)
automatic repairs would be nice, but lets save that until we know that this
ticket works?
> Avoid repairing already-repaired data by default
> ------------------------------------------------
>
> Key: CASSANDRA-5351
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5351
> Project: Cassandra
> Issue Type: Task
> Components: Core
> Reporter: Jonathan Ellis
> Assignee: Lyuben Todorov
> Labels: repair
> Fix For: 2.1
>
>
> Repair has always built its merkle tree from all the data in a columnfamily,
> which is guaranteed to work but is inefficient.
> We can improve this by remembering which sstables have already been
> successfully repaired, and only repairing sstables new since the last repair.
> (This automatically makes CASSANDRA-3362 much less of a problem too.)
> The tricky part is, compaction will (if not taught otherwise) mix repaired
> data together with non-repaired. So we should segregate unrepaired sstables
> from the repaired ones.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
