Re: Tomcat Clustering trouble when starting up under high load

2009-03-11 Thread Mikel Ibiricu
Hi Rainer, thanks for your response. I tried the config you suggested. It suppose the way of configuring those parameters is as you said, but skipping manager. from the begining, as when I tried as you said, I got this in the catalina. log: 10-mar-2009 15:58:19

Re: Tomcat Clustering trouble when starting up under high load

2009-03-11 Thread Filip Hanik - Dev Lists
hi Mikel, when setting a property on the Manager you omit the manager. prefix, just as you stated. To do thread dumps with JDK 1.5 under windows, you can use the tanuki service wrapper http://people.apache.org/~fhanik/wrapper.html The tomcat team might have added that feature to the Tomcat

Re: Tomcat Clustering trouble when starting up under high load

2009-03-09 Thread Mikel Ibiricu
Hi Filip Thanks for your response. We have been testing some modifications on our config, specially focusing in what you told us about limiting stateTransferTimeout, which we have limited to 180 seconds now. Actually, it does not get stuck, in the worst case, it only starts without replicating

Re: Tomcat Clustering trouble when starting up under high load

2009-03-09 Thread Rainer Jung
On 09.03.2009 09:24, Mikel Ibiricu wrote: So, It works OK but when starting up one of the nodes with over 500 sessions alive in the other, it doesn't replicate anything. We assume that it would not be able to replicate everything... but why it does either replicate everything or nothing? If it's

Re: Tomcat Clustering trouble when starting up under high load

2009-01-30 Thread Filip Hanik - Dev Lists
state transfer timeout -1 is not a good setting. one should prefer to timeout rather than getting stuck, even if the timeout means we didn't get everything you may also try the backup manager, which does the state transfer in a bit smarter manner. when your system is stuck, then thread

Tomcat Clustering trouble when starting up under high load

2009-01-29 Thread Mikel Ibiricu
Hello all I´m Mikel, and me and my workmates have already been a while testing our environment in order to establish a in memory session replication cluster for our servers. The thing is that our servers are often loaded with up to a thousand (1000) and more sessions (and now, we have three