[
https://issues.apache.org/jira/browse/AVRO-291?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Doug Cutting resolved AVRO-291.
-------------------------------
Resolution: Fixed
Fix Version/s: 1.3.0
Assignee: Eric Evans
Hadoop Flags: [Reviewed]
I just committed this. Thanks, Eric!
I changed it to, instead of catching the IOException and rethrowing it as a
RuntimeException, to simply permit IOException to be thrown. Most RPC code
must already handle IOException.
> SocketServer does not set NODELAY
> ---------------------------------
>
> Key: AVRO-291
> URL: https://issues.apache.org/jira/browse/AVRO-291
> Project: Avro
> Issue Type: Bug
> Components: java
> Affects Versions: 1.3.0
> Environment: Linux, OpenJDK 1.6
> Reporter: Eric Evans
> Assignee: Eric Evans
> Fix For: 1.3.0
>
> Attachments: 0001-enable-SO_NODELAY-on-SocketTransceiver.patch
>
>
> As discussed on the list
> (http://www.mail-archive.com/avro-u...@hadoop.apache.org/msg00098.html),
> without NODELAY, it's possible for RPC transactions to be overly latent, (on
> Linux, 40ms worth of ack timeout is added in each direction when message
> sizes are small).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
