[
https://issues.apache.org/jira/browse/CASSANDRA-6689?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13914452#comment-13914452
]
Benedict commented on CASSANDRA-6689:
-------------------------------------
Okay, pushed a patch
[here|https://github.com/belliottsmith/cassandra/tree/offheap1b] which includes
all of your branch, is up-to-date with 2.1, and also:
bq. In NBQV, drainTo does not return the correct number of inserted items.
I've fixed this, and also merged NQB.drainTo() to some extent with
NBQV.drainTo, to optimise it a little as well, and added some long tests to
NonBlockingQueueTest to cover it
bq. Make OpOrder.Group and RefAction implement AutoClosable to be able to do
try-with-resources?
I've introduced this change, and I've updated all of the try/finally blocks to
use it. Much neater.
Things I haven't touched:
bq. AtomicReferenceArrayUpdater.shift(), really?
I'm really not sure what we can improve with this?
bq. I flush ~40MB/s with HeapSlabPool but only ~15MB/s with OffHeapPool.
I think we should split this into another ticket. I suspect we might be able to
improve this by introducing an extension of DataOutput which can accept a
ByteBuffer.
bq. GC metrics! Time spent, amount of data reallocated etc.
Also think this deserves its own ticket.
> Partially Off Heap Memtables
> ----------------------------
>
> Key: CASSANDRA-6689
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6689
> Project: Cassandra
> Issue Type: New Feature
> Components: Core
> Reporter: Benedict
> Assignee: Benedict
> Fix For: 2.1 beta2
>
>
> Move the contents of ByteBuffers off-heap for records written to a memtable.
> (See comments for details)
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
