[
https://issues.apache.org/jira/browse/CASSANDRA-9738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14742716#comment-14742716
]
Benedict commented on CASSANDRA-9738:
-------------------------------------
bq. In hindsight wholesale replacing the existing key cache was not a good
incremental way to effect this change in a low risk way
So, looking at the latest performance comparisons, it doesn't look like the
benefit is so uniformly profound as to be worth rushing this out. It looks like
whatever was particularly affecting 3.0 in the early benchmarks (that notably
wasn't affecting prior versions) has been ironed out.
If the code is still being majorly revised so close to RC, we should consider
pushing this back, and if incorporated before GA at least consider making it
optional. It's turning into a patch of really significant size, and we already
have a great deal of risk associated with this release.
> Migrate key-cache to be fully off-heap
> --------------------------------------
>
> Key: CASSANDRA-9738
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9738
> Project: Cassandra
> Issue Type: Sub-task
> Reporter: Robert Stupp
> Assignee: Robert Stupp
> Fix For: 3.0.0 rc1
>
>
> Key cache still uses a concurrent map on-heap. This could go to off-heap and
> feels doable now after CASSANDRA-8099.
> Evaluation should be done in advance based on a POC to prove that pure
> off-heap counter cache buys a performance and/or gc-pressure improvement.
> In theory, elimination of on-heap management of the map should buy us some
> benefit.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
