[jboss-jira] [JBoss JIRA] Commented: (JGRP-506) loopback should use thread pool to deliver message up the stack

Fri May 11 08:58:54 EDT 2007

    [ http://jira.jboss.com/jira/browse/JGRP-506?page=comments#action_12362052 ] 

Michael Newcomb commented on JGRP-506:
--------------------------------------

Say you have a long running task (relatively long running) that you want the group to do. If you use loopback, the local node has to complete the operation before it is sent to the group. If each operation on a node in the cluster takes a similar time, say 1s, to complete, the entire operation will now take 2s because it first operates on the local node (taking 1s to complete), then sends out the task where every node executes the task and it takes 1s more (for a total of 2s).

I thought loopback was mainly to save network overhead.

> loopback should use thread pool to deliver message up the stack
> ---------------------------------------------------------------
>
>                 Key: JGRP-506
>                 URL: http://jira.jboss.com/jira/browse/JGRP-506
>             Project: JGroups
>          Issue Type: Bug
>            Reporter: Michael Newcomb
>         Assigned To: Bela Ban
>
> If loopback is true, the thread calling down() also delivers the message up().

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira