Bug 51736 - Make rpcTimeout configurable in BackupManager.
Summary: Make rpcTimeout configurable in BackupManager.
Alias: None
Product: Tomcat 6
Classification: Unclassified
Component: Cluster (show other bugs)
Version: unspecified
Hardware: All All
: P2 enhancement (vote)
Target Milestone: default
Assignee: Tomcat Developers Mailing List
Depends on:
Reported: 2011-08-29 10:26 UTC by Keiichi Fujino
Modified: 2011-08-31 02:03 UTC (History)
0 users


Note You need to log in before you can comment on or make changes to this bug.
Description Keiichi Fujino 2011-08-29 10:26:09 UTC
Default rpcTimeout that BackupManager sets to AbstractReplicatedMap is 15000 msec.
This is not configurable.

When receiving MapMessage.MSG_START (and after a new member is added), 
publishEntryInfo is called for all sessions where primary and backup node are not included. 
And, the backup node is set to the session.

When only one Tomcat starts in cluster, publishEntryInfo will be called for all sessions. 
As a result, the RPC message has the possibility of the timeout. 

When acquiring session from restarting Tomcat, 
warning of "Unable to retrieve remote object for key:XXX" is logged.
And the session cannot be continued. 

I make rpcTimeout configurable in BackupManager.
Additionally, if RPC message timeouted in AbstractReplicatedMap#broadcast, add warning message.

Best Regards.
Comment 1 Keiichi Fujino 2011-08-29 11:17:32 UTC
Fixed in 7.0.x and will be in 7.0.21 onwards.
Proposed for 6.0.x.
Comment 2 Keiichi Fujino 2011-08-31 02:03:53 UTC
Fixed in 6.0.x and will be in 6.0.34 onwards.