Skip to main content

RMI/ACC Fail-over failing

Please note these java.net forums are being decommissioned and use the new and improved forums at https://community.oracle.com/community/java.
1 reply [Last post]
smithbr4
Offline
Joined: 2011-07-28

We currently are running Glassfish 3.1.2.2 in a 5 node cluster(for scalability) we have an application client launched via the Javaweb Start. This client connects to the glassfish server using RMI via the ACC. Everything runs really great until we run into one of these bugs(GLASSFISH-16217, GLASSFISH-18351) which causes one of the nodes to deadlock. When the node deadlocks the ACC is not smart enough to remove the deadlocked node from the list, the client freezes and becomes un-usable. Basically we have this sports car that goes really fast but we have no seatbelts or any kind of safety features. Is there anyway to change the way the ACC determines node failure so we can call our own heartbeat service or something? This is really frustrating. Any feedback would be greatly appreciated.

Thanks,
Brody Smith

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
preston001
Offline
Joined: 2008-11-19

In reading the GlassFish_v3.1_GMS_One_Pager it seems like a FailureNotificationSignal could be generated which would then cause the removal of the node from the IIOPEndpoints.

Can anyone else confirm this would work or is the right approach?