Re: [jgroups-dev] Problem with FD_SOCK failure detection

yes, this is reproducible. 
FD_SOCK start() method is not called as JChannel.connect call wasn't
succeeded because it was trying to JOIN killed member(co-ordinator)

there are three nodes in cluster. (member-89,member-101,member-88,member-87)
we are getting lot of warnings mentioned below. it is trying to contact
member-89(which is killed) 
[pool-3-thread-1] WARN  o.jgroups.protocols.pbcast.GMS -
JOIN(member-87-22015) sent to member-89-55050 timed out (after 7000 ms), on
try 2852

On discovery request from member-87: member-88 is returning view
[member-89-55050|10]. So it is always sending JOIN request to
member-89(killed).

member-88,89,101 has view {member-89,member-101,member-88,member-87}
member-87 does't have any view. 
There is trace log on member-87 which i believe is view change message but
it doesn't get passed up.
"TRACE o.j.protocols.pbcast.NAKACK2   - member-88-44628: message
member-8955050::9 was added to queue (not yet server)"

Scenario is: let say 2  A,B,C is in cluster new node D try to join and at
the same time A goes down. than C has already closed the socket with A
gracefully, assuming D will create the socket with A . D's FD_SOCK pinger
thread is not started. ( i am not seeing the msg "pinger_thread started" in
logs)

on B and C view contain {A,B,C,D}, D didn't received the cluster view
message.

==
We are not using combo of FD_SOCK with FD/FD_ALL. we are using only in
FD_SOCK. But i think this issue will come with FD too and FD also detects
right neighbor and initialization of these two protocol is same.


==
I have changed determinePingDest() to use detect left neighbor. Is there any
thing else that's needs to change also





--
View this message in context: http://jgroups.1086181.n5.nabble.com/Problem-with-FD-SOCK-failure-detection-tp9847p9850.html
Sent from the JGroups - Dev mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Introducing Performance Central, a new site from SourceForge and 
AppDynamics. Performance Central is your source for news, insights, 
analysis and resources for efficient Application Performance Management. 
Visit us today!
http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk
_______________________________________________
Javagroups-development mailing list