[
https://issues.apache.org/jira/browse/HDFS-17875?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ivan Andika updated HDFS-17875:
-------------------------------
Description:
[~XiChen] found that AtomicFileOutputStream write is slow because the
`write(byte b[], int off, int len)` uses FilterOutputStream implementation
which calls write(byte b) for each byte.
We should override this API and forward it to the underlying OutputStream
instead.
was:
[~XiChen] found that AtomicFileOutputStream write is slow because the `
write(byte b[], int off, int len)` uses FilterOutputStream implementation which
calls write(byte b) for each byte.
We should override this API and forward it to the underlying OutputStream
instead.
> AtomicFileOutputStream should override FilterOutputStream batch write API
> -------------------------------------------------------------------------
>
> Key: HDFS-17875
> URL: https://issues.apache.org/jira/browse/HDFS-17875
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Ivan Andika
> Priority: Major
>
> [~XiChen] found that AtomicFileOutputStream write is slow because the
> `write(byte b[], int off, int len)` uses FilterOutputStream implementation
> which calls write(byte b) for each byte.
>
> We should override this API and forward it to the underlying OutputStream
> instead.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
