Re: [ERROR] NAKACK.handleXmitReq()

Gianluca Puggelli wrote:
> Bela,
> at the moment I'm around 800 million of messages without any
> problem, with the protocol stack that you sent with the previous
> message (tcp_nio).

Good. I'm currently at 1'874'000'000 messages received (out of 6 
billion), memory is stable.


> One hour ago I attached the jconsole to java processes and now I
> discovered that there are 39 live threads in the coordinator while
> 34 are in the other two members.
>
> But in one hour the coordinator created ~550 threads while the other
> members only ~50. Do you know why ?

The coordinator has some additional tasks to do, compared to regular 
(non-coordinator) members, e.g.

    * Merge: coord sends out a merge multicast every 20-100secs
    * FD ((everyone though) sends out are-you-alive messages)
    * With TCP_NIO, idle threads will be killed off, and started at a
      later time again if needed (we use PooledExecutor)
    * Timer

The peak number of threads is always higher on the coord, but the number 
of live threads should be about the same

-- 
Bela Ban
Lead JGroups / Manager JBoss Clustering Group
JBoss - a division of Red Hat

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642