[CASSANDRA-1878] Minimize Key Cache Invalidation by Compaction

• Type: Improvement
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 0.6.9, 0.7.0 rc 3
• Assignee: Jonathan Ellis
• Reporter: Tom Hobbs

Authors

Jonathan Ellis

Reviewers

T Jake Luciani

• Created: 18/Dec/10 00:15
• Updated: 16/Apr/19 09:33
• Resolved: 20/Dec/10 04:45

Description

Currently, compactions invalidate key cache entries that pointed to the SSTables that were compacted. This results in a sudden increase in the number of seeks necessary for reads immediately after compaction. When writing out a new SSTable after a compaction, it seems like it should be possible to store a list of keys that are currently cached but would be invalidated along with their new position in the new SSTable. Matt Dennis also seems to think this would be relatively easy to do.