[ https://issues.apache.org/jira/browse/CASSANDRA-11452?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15240956#comment-15240956 ]
Branimir Lambov commented on CASSANDRA-11452: --------------------------------------------- I'm sorry, I don't see how this helps. Once both the hot key and its collision are in the main area (this check is not enough to guarantee that won't happen though it probably manages to do so for this specific test), this path is no longer triggered. I think we should be looking for a way to eject an offending entry after it has entered. I think an ideal test would verify that CLASH is among the cached keys before "// Now run a repeating sequence ...", but not there after the loop has finished. Did you run traces with candidate preference on equality? Is it still bad? > Cache implementation using LIRS eviction for in-process page cache > ------------------------------------------------------------------ > > Key: CASSANDRA-11452 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11452 > Project: Cassandra > Issue Type: Improvement > Components: Local Write-Read Paths > Reporter: Branimir Lambov > Assignee: Branimir Lambov > > Following up from CASSANDRA-5863, to make best use of caching and to avoid > having to explicitly marking compaction accesses as non-cacheable, we need a > cache implementation that uses an eviction algorithm that can better handle > non-recurring accesses. -- This message was sent by Atlassian JIRA (v6.3.4#6332)