[
https://issues.apache.org/jira/browse/HBASE-15506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15212042#comment-15212042
]
Vladimir Rodionov commented on HBASE-15506:
-------------------------------------------
Yes, I tried with CMS. I do observe reduced memory allocation, especially in
old gen space. It seems, that these buffers get allocated directly in old gen
(tenured - CMS), because their default size exceeds the max size for young gen.
Even if there are no obvious performance benefits, allocating less garbage
allow system to run more smoothly and have less hiccups (promotion failure,
full gc) - that is my opinion.
> FSDataOutputStream.write() allocates new byte buffer on each operation
> ----------------------------------------------------------------------
>
> Key: HBASE-15506
> URL: https://issues.apache.org/jira/browse/HBASE-15506
> Project: HBase
> Issue Type: Improvement
> Reporter: Vladimir Rodionov
> Assignee: Vladimir Rodionov
>
> Deep inside stack trace in DFSOutputStream.createPacket.
> This should be opened in HDFS. This JIRA is to track HDFS work.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
