[
https://issues.apache.org/jira/browse/HBASE-19282?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16289067#comment-16289067
]
Anastasia Braginsky commented on HBASE-19282:
---------------------------------------------
Generally, to use chunks smaller than 2MB (when chunk index is really smaller
than 2MB) is a great idea. However, chunks must be allocated via ChunkCreator.
What is the difference whether they are allocated "from MSLAB" or as an
"on-heap ByteBuffer"? When working on-heap the MSLAB's chunks are after all
on-heap ByteBuffers.
We can create a "variable size chunks" as extension and generalization to the
jumbo chunks we have done. However, this was not the purpose of this JIRA and
shouldn't been done as part of this JIRA. This JIRA is about setting the user
configuration for CCM. I'll open another JIRA about "variable size chunks" to
support the case when a single CCM index chunk is smaller than 2MB. This would
require multiple changes.
> CellChunkMap Benchmarking and User Interface
> --------------------------------------------
>
> Key: HBASE-19282
> URL: https://issues.apache.org/jira/browse/HBASE-19282
> Project: HBase
> Issue Type: Sub-task
> Reporter: Anastasia Braginsky
> Attachments: CCM Benchmarking.pdf
>
>
> We have made some experiments how working with CellChunkMap (CCM) influences
> the performance when running on-heap and off-heap. Based on those results it
> is suggested to tie the MSLAB usage (off-heap or on-heap) with CCM index
> usage.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
