cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ariel Weisberg (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-9738) Migrate key-cache to be fully off-heap
Date Tue, 08 Sep 2015 20:54:46 GMT


Ariel Weisberg commented on CASSANDRA-9738:

Just going off of coverage from "ant test" not including your work today. I have had some
bad luck recently with emma claiming stuff is not covered. Hopefully it's accurate in these

* Serializers.LegacyClusteringPrefixSerializer.deserialize isn't tested. Is that blocked by
* IndexInfo.skip isn't labeled as covered and neither is serializedSize
* RowIndexEntry.equals is not covered, neither is hashCode, same for RIE.IndexedEntry
* RowIndexEntry.Serializer.serialize doesn't test manually serializing the old version
* RowIndexEntry.IndexedEntry constructor doesn't test loading the old version
* RowIndexEntry.IndexedEntry.indexInfo doesn't test the old version, also the branch at the
end where it stores the offset it happens to know about after finishing
* RowIndexEntry.IndexedEntry.indexInfo does it have to do that loop for already discovered
offsets? Can you store that value and not loop?
* RIE.IndexedEntry.indexCount() could store the offset as a constant, also firstIndexInfoOffset
* RIE.IndexedEntry.promotedSize for the old version not covered
* Slice.Serializer.skip() is not tested/covered, neither is serialized size or part of deserialize,
but you didn't change the latter two
* [RowIndexEntyrTest has commented out code|]

Reviewing test contents now. Also going to update and look at your latest. 

> Migrate key-cache to be fully off-heap
> --------------------------------------
>                 Key: CASSANDRA-9738
>                 URL:
>             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

View raw message