This looks like a bug.
I tried running two instances on my machine and I dont see this though.
Could you tell me if you downloaded before yesterday? We had an outdated
build before that on the shoal pages.
regards
Shreedhar
Kapucin wrote:
> Hello,
>
> I'm interested in master fail over functionality. Your framework seems
> to provide it, but I couldn't get it working (tried ApplicationServer
> example and ClusterManager app).
> I ran two instances (core type) on two different hosts. At start-up
> one becomes Master the other acknowledges it. When I kill one instance
> the other sees it but does NOT try to become a master (isMaster: false).
>
> [#|2007-05-11T12:28:20.929-0400|FINER|Shoal|JxtaMgmt|_ThreadID=17;_ThreadName=FailureVerifier
> Thread;ClassName=MasterNode;MethodName=isMaster;|isMaster :false
> MasterAssigned :true View Size :2|#]
>
> When the original master node is restarted it keeps sending a system
> advertisement to the other node non-stop every 3 seconds (it did not
> do it before crash):
>
> [#|2007-05-11T12:42:34.580-0400|INFO|Shoal|JxtaMgmt|_ThreadID=13;_ThreadName=pool-2-thread-1;ClassName=HealthMonitor;MethodName=process;urn:jxta:uuid-F310105316A8469284C7
> B474506B9A4B18E1E08DB2D34D26870F9400BA1D4CE703;DEF;|probing.id|#]
> [#|2007-05-11T12:42:34.580-0400|FINER|Shoal|JxtaMgmt|_ThreadID=13;_ThreadName=pool-2-thread-1;ClassName=MasterNode;MethodName=isMaster;|isMaster
> :true MasterAssigned :true View Size :1|#]
> [#|2007-05-11T12:42:34.596-0400|FINER|Shoal|JxtaMgmt|_ThreadID=13
> ;_ThreadName=pool-2-thread-1;ClassName=MasterNode;MethodName=createNodeQuery;|Created
> a Node Query Message |#]
> [#|2007-05-11T12:42:34.596-0400|FINER|Shoal|JxtaMgmt|_ThreadID=13;_ThreadName=pool-2-thread-1;ClassName=MasterNode;MethodName=send;|Unicasting
> Message to
> :nullID=urn:jxta:uuid-F310105316A8469284C7B474506B9A4B18E1E08DB2D34D26870F9400BA1D4CE703|#]
>
>
>
> On which the other node prints:
>
> [#|2007-05-11T12:39:57.775-0400|FINER|Shoal|JxtaMgmt|_ThreadID=18;_ThreadName=pool-1-thread-5;ClassName=MasterNode;MethodName=processNodeAdvertisement
> ;|Received a System advertisment Name :XXX|#]
> [#|2007-05-11T12:39:57.775-0400|FINER|Shoal|JxtaMgmt|_ThreadID=18;_ThreadName=pool-1-thread-5;ClassName=MasterNode;MethodName=isMaster;|isMaster
> :false MasterAssigned :true View Size :1|#]
>
> Is it expected behavior or am I missing something? Do I need to change
> anything in the configuration to enable master fail over?
>
> Regards