Re: [javagroups-users] Strange condition

[Bela B. <be...@ya...>]

A deadlock doesn't depend on the stack config, just on whether you;re 
making synchronous cluster method calls.
Next time this blocks, take a stack dump and post it. Best is to have 
stack dumps of *all* members

Mark K wrote:
>> Hmm, are you doing anything like a synchronous method call across the 
>> cluster in your DistributedHashListener ?
>
> Nope! Not at all, the DistributedHashListener, just prints out to 
> console! It's defintely not the source of deadlock here!
>
> The error is hard to replicate, a stacktrace might be difficult.
>
> Distributed Deadlock sounds like a cool explanation, now  is it 
> possible to have distributed deadlock with these elements in protocol 
> stack, theoretically or otherwise?
>
> <TCP start_port="9888" loopback="false" />
> <TCPPING timeout="10000" initial_hosts="192.168.2.91[9888]" 
> port_range="5" num_initial_members="4"/>
> <FD timeout="10000" max_tries="4"/>
> <VERIFY_SUSPECT timeout="5000" down_thread="false" up_thread="false"/>
> <pbcast.NAKACK gc_lag="100" retransmit_timeout="600,1200,2400,4800"/>
> <pbcast.STABLE stability_delay="1000" desired_avg_gossip="20000" 
> down_thread="false" max_bytes="0" up_thread="false"/>
> <pbcast.GMS print_local_addr="true" join_timeout="5000" 
> join_retry_timeout="2000" shun="true"/>
> <pbcast.STATE_TRANSFER />
>

-- 
Bela Ban
Lead JGroups / Manager JBoss Clustering Group
callto://belaban