Re: [jgroups-dev] FD_ALL Question

David Forget wrote:
> Hi, I am looking at a potential issue with JGroups 2.8GA related to 
> FD_ALL.
> As I understand it every config:intervale FD_ALL is multicasting a 
> HeartBeat
> and verify if HB from all other cluster members were received within the
> config:timeout. If a node did not send HB or has not been received within
> the config:timeout a FD_ALL:SUSPECT message is multicast, nodes that
> receives the message propagate it up on JGroups stack, when it reach 
> FD_ALL
> protocol the event is passed UP and DOWN. Why passing the event DOWN in
> FD_ALL.java Ligne 172 ?

There might be other protocols which are interested in knowing that a 
cluster node was suspected, and want to handle the SUSPECT event.

Currently, no protocol below FD_ALL (in the default config) does that, 
but that doesn't mean it won't get handled in the future.

Are you concerned about the cost ? Shouldn't cost much, as this is just 
a spurious down call.

-- 
Bela Ban
Lead JGroups / Clustering Team
JBoss

------------------------------------------------------------------------------
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store 
http://p.sf.net/sfu/nokia-dev2dev
_______________________________________________
Javagroups-development mailing list