[
https://issues.apache.org/jira/browse/HDFS-10326?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15373530#comment-15373530
]
Arpit Agarwal commented on HDFS-10326:
--------------------------------------
{{DEFAULT_DATA_SOCKET_SIZE}} is not used as a hard-coded default after
HDFS-8829 and HDFS-9259. The following settings control socket buffers for data
pipeline connections. All are initialized to DEFAULT_DATA_SOCKET_SIZE.
# dfs.datanode.transfer.socket.send.buffer.size
# dfs.datanode.transfer.socket.recv.buffer.size
# dfs.client.socket.send.buffer.size
We could change DEFAULT_DATA_SOCKET_SIZE to enable TCP window auto-tuning by
default. Linux kernel 2.6.6+ supports send and receive window auto-tuning which
covers Redhat/Centos 5+ and Ubuntu 4+.
> Disable setting tcp socket send/receive buffers for write pipelines
> -------------------------------------------------------------------
>
> Key: HDFS-10326
> URL: https://issues.apache.org/jira/browse/HDFS-10326
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode, hdfs
> Affects Versions: 2.6.0
> Reporter: Daryn Sharp
> Assignee: Daryn Sharp
>
> The DataStreamer and the Datanode use a hardcoded
> DEFAULT_DATA_SOCKET_SIZE=128K for the send and receive buffers of a write
> pipeline. Explicitly setting tcp buffer sizes disables tcp stack
> auto-tuning.
> The hardcoded value will saturate a 1Gb with 1ms RTT. 105Mbs at 10ms.
> Paltry 11Mbs over a 100ms long haul. 10Gb networks are underutilized.
> There should either be a configuration to completely disable setting the
> buffers, or the the setReceiveBuffer and setSendBuffer should be removed
> entirely.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
