[
https://issues.apache.org/jira/browse/HBASE-9461?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13766457#comment-13766457
]
Nicolas Liochon commented on HBASE-9461:
----------------------------------------
Just asking, if you find the time. Some comments describing the intention (what
this class is supposed to do) would be great. For example, the delayed calls,
undelayed calls with delayed response are not trivial (are they used?). It's
very difficult to understand what these classes are supposed to do and why.
Or this low level, with a specific case with '==1'
{code}
synchronized (call.connection.responseQueue) {
call.connection.responseQueue.addLast(call);
if (call.connection.responseQueue.size() == 1) {
doRegister = !processResponse(call.connection.responseQueue, false);
}
}
{code}
Basically, I wanted to try a Netty based implementation, but there are so many
corner cases that even estimating the time needed for a hacked implementation
is difficult....
> Some doc and cleanup in RPCServer
> ---------------------------------
>
> Key: HBASE-9461
> URL: https://issues.apache.org/jira/browse/HBASE-9461
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Assignee: stack
> Attachments: 9461.txt, 9461v2.txt, ipc2.ucls
>
>
> RPC is a dog to follow. I want to do buffer pooling for reading requests but
> its tough drawing the diagram of who is doing what when. HBASE-8884 seems to
> have made it more involved still. This issue is about doing a bit of
> untangling.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
