Member propagates non-existing coordinator ------------------------------------------ Key: JGRP-679 URL: http://jira.jboss.com/jira/browse/JGRP-679 Project: JGroups Issue Type: Bug Affects Versions: 2.2.8 Environment: Oracle Application Server Cluster running application using JGroups Reporter: Peter Fuchs Assigned To: Bela Ban Priority: Critical Attachments: cms_application.log.gz Our application is running on a clustered application server environment (Oracle AS 10gR2) with 9 application instances in total. We are using JGroups for keeping the object caches of all application instances in sync. From time to time we encounter the problem that our application cannot be started and seems to "hang". After setting the JGroups logger to DEBUG we could see that it was JGroups that was trying to connect to a coordinator that did not exist anymore. After some investigation we found out that there was one running instance propagating a coordinator that did not exist any longer. The only solution in that case was to restart the instance that was propagating the wrong coordinator address. As the version we currently use (2.2.8) is quite old I checked Jira for fixes regarding this issue but could not really identify this specific behaviour (maybe due to my poor knowledge about the internals of JGroups). We are planning to use JGroups V2.4.1(SP4) in our future software releases (as we are bound to JDK 1.4.2) but would like to know if this issue has been fixed in that version. Please find attached an example log file excerpt.

Member propagates non-existing coordinator ------------------------------------------ Key: JGRP-679 URL: http://jira.jboss.com/jira/browse/JGRP-679 Project: JGroups Issue Type: Bug Affects Versions: 2.2.8 Environment: Oracle Application Server Cluster running application using JGroups Reporter: Peter Fuchs Assigned To: Bela Ban Priority: Critical Attachments: cms_application.log.gz Our application is running on a clustered application server environment (Oracle AS 10gR2) with 9 application instances in total. We are using JGroups for keeping the object caches of all application instances in sync. From time to time we encounter the problem that our application cannot be started and seems to "hang". After setting the JGroups logger to DEBUG we could see that it was JGroups that was trying to connect to a coordinator that did not exist anymore. After some investigation we found out that there was one running instance propagating a coordinator that did not exist any longer. The only solution in that case was to restart the instance that was propagating the wrong coordinator address. As the version we currently use (2.2.8) is quite old I checked Jira for fixes regarding this issue but could not really identify this specific behaviour (maybe due to my poor knowledge about the internals of JGroups). We are planning to use JGroups V2.4.1(SP4) in our future software releases (as we are bound to JDK 1.4.2) but would like to know if this issue has been fixed in that version. Please find attached an example log file excerpt.