cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pavel Yaskevich (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-6689) Partially Off Heap Memtables
Date Sat, 22 Mar 2014 20:20:46 GMT


Pavel Yaskevich commented on CASSANDRA-6689:

bq. My understanding (not my policy) is that this is explicitly not a default supported option
(in the current state of affairs) because of the expectation of increased heap fragmentation
and GC death spiral, which is why I'm happy with the ugly name for that one option.

So we can go both ways - remove it completely from the allocation type enum and don't support
it or leave it in place and add it to documentation so users can make informed decision weather
to use it or not meanwhile default would still be "heap_slab" or similar name.

bq. In this case what is to be gained? Seems likely it would only increase the need for old
gen collections, as we wouldn't be able to do any reuse?

It could perform the same function as previous heap slab, for par new it doesn't make much
different but if slab slips to the old get then it would create less fragmentation comparing
to individual buffers getting promoted (plus we can pack some fresh allocations inside), but
that's just an idea and isn't critical for this ticket...

> Partially Off Heap Memtables
> ----------------------------
>                 Key: CASSANDRA-6689
>                 URL:
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Benedict
>            Assignee: Benedict
>              Labels: performance
>             Fix For: 2.1 beta2
>         Attachments: CASSANDRA-6689-final-changes.patch, CASSANDRA-6689-small-changes.patch
> Move the contents of ByteBuffers off-heap for records written to a memtable.
> (See comments for details)

This message was sent by Atlassian JIRA

View raw message