Details
-
Type:
Bug Fix
-
Status:
Closed
-
Priority:
Major
-
Resolution: Complete
-
Affects Version/s: None
-
Fix Version/s: 2.4
-
Component/s: Development
-
Security Level: Public (Public: Anyone can view)
-
Labels:
-
Similar issues:
KULRICE-13491 Test failover of Rice components KULRICE-2209KSB failover implementation is inadequate KULRICE-3335Failover is not working properly in the KSB in the case of a 404 coming back from a service KULRICE-9151Document in the guide on clustering that session failover should be implemented for full reliability KULRICE-1531Embedded client needs to have identical service name for OSCacheNotificationService topic so that it can recieve cache messages KULRICE-9032cancelling a document with client side validation disabled KULRICE-6790ClassCastException when background KSB message processing invokes kimCacheAdminService (embedded mode) KULRICE-1786Is note topic functionality broken? KULRICE-1012KSB Message queue blowing up KULRICE-7633BusClientFailoverProxy needs to be handling org.apache.cxf.transport.http.HTTPException as a failover exception -
Rice Module:KSB
-
KAI Review Status:Not Required
-
KTI Review Status:Not Required
-
Code Review Status:Not Required
-
Contributing Institution:Indiana Univ
-
Include in Release Notes?:Yes
Description
When the KSB tries to send out a topic style message to each server that is online (e.g. a cache flush message) it should not use the BusClientFailureProxy. If a topic message fails to be sent to a specific node the BusClientFailureProxy makes it retry the message to an alternate endpoint. In the case where that topic message is something like a cache flush that means the caches end up in a bad state on one or more machines in a cluster. The fix to this issue is to make sure the service is not wrapped with the failure proxy for topic messages so if the call fails it will retry it to the same node.
Contribution complete.