Jonathan Ellis created CASSANDRA-5449:
-----------------------------------------
Summary: Make sstable compacting status un/marking less error-prone
Key: CASSANDRA-5449
URL: https://issues.apache.org/jira/browse/CASSANDRA-5449
Project: Cassandra
Issue Type: Bug
Reporter: Jonathan Ellis
Priority: Minor
As demonstrated by a393072aac6292412fc465d207c411c4b6b69e0b, it's easy to
introduce regressions where we don't unmark the same tables we marked. This is
primarily because the marking and unmarking are usually done by separate
methods. (The opposite problem is also possible -- performAllSSTableOperation
unmarks compacting, and so does CompactionTask.execute, which can be part of a
pASOp via the scrub path.)
I suggest making markCompacting return a callable that will wrap the
caller-provided code in a try/finally to centralize this.
--
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
