[
https://issues.apache.org/jira/browse/HBASE-15555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15216119#comment-15216119
]
Anoop Sam John commented on HBASE-15555:
----------------------------------------
Ya it is like 14490 change. The BBBPool gives us DBBs and we read into it
directly. Then the Codec decoder makes cells directly out of these DBBs. Later
on at memstore level the cells will be copied to MSLAB area (we do have an
offheap MSLAB implementation also). This reduces the #GCs but the avg pause
time is increased. Need to investigate more here specially wrt GC setting and
tuning. Will post back our findings soon.
> Memory Management
> -----------------
>
> Key: HBASE-15555
> URL: https://issues.apache.org/jira/browse/HBASE-15555
> Project: HBase
> Issue Type: Umbrella
> Reporter: stack
> Attachments: download.svg
>
>
> Umbrella issue on memory management. One-stop shop to learn how we're doing
> it, tenets, and work to be done. Collect it all in one place so we are
> aligned rather than each trying to do it for themselves.
> Subtasks include:
> + Map of memory allocation and recycling landscape
> + Where are we allocating when we could be recycling
> + MSLAB and G1GC don't go together?
> + Enable MSLAB pool always?
> + With offheap allocations, less GC but it runs for longer?
> + Better heuristics recycling: e.g. HBASE-15525 OutOfMemory could occur when
> using BoundedByteBufferPool during RPC bursts
> + See if can do DFSClient buffer reuse: HBASE-15506
> FSDataOutputStream.write() allocates new byte buffer on each operation
> What else?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
