[
https://issues.apache.org/jira/browse/CASSANDRA-7882?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14166426#comment-14166426
]
Jay Patel commented on CASSANDRA-7882:
--------------------------------------
[~benedict], we've done initial testing with 3K tables, running for a day &
looks good so far.
Should we keep the start region size configurable, instead of hard-coded 8kb?
This will allow fine tuning specific use case, for instance - If payload is
always around 400 to 500kb, user can have start size as 512kb instead of
scaling from 8kb. Actually, we've one use case like that.
Let me know your thoughts or I'll submit a final patch with the todos completed.
> Memtable slab allocation should scale logarithmically to improve occupancy
> rate
> -------------------------------------------------------------------------------
>
> Key: CASSANDRA-7882
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7882
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Jay Patel
> Assignee: Jay Patel
> Labels: performance
> Fix For: 2.1.1
>
> Attachments: trunk-7882.txt
>
>
> CASSANDRA-5935 allows option to disable region-based allocation for on-heap
> memtables but there is no option to disable it for off-heap memtables
> (memtable_allocation_type: offheap_objects).
> Disabling region-based allocation will allow us to pack more tables in the
> schema since minimum of 1MB region won't be allocated per table. Downside can
> be more fragmentation which should be controllable by using better allocator
> like JEMalloc.
> How about below option in yaml?:
> memtable_allocation_type: unslabbed_offheap_objects
> Thanks.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
