I have succesfully installed SJSAS 8.1 on Solaris 9 (sparc architecture). I am able to fully configure and operate my application server. However, I have a similar problem to the one you signaled above. Except I do not fully understand what you meant by "abruptly". In my case, the SJSAS is installed on a Sun Sparc Server. I run the start-domain and start-node-agent commands either from a telnet console, or through a hyper-terminal on a machine directly connected to the server, or using WRQ Reflection X. I ALL cases, both the domain and the node-agent stop when I close wichever tool I used to start it. (i.e. as soon as I close my telnet connection, the SJSAS domain and node-agent are stopped). I know that maybe the problem lies in the fact that Solaris sees the SJSAS asadmin as a child process of the console I am running, hence they both die when the console dies. Still I wonder if there is any way of making de domain and node-agent keep alive without the need of a console being open.
Any help I can get would be most appreciated.
Thank you.
Miguel醤gel.
hi,
Do you know the exact version of the appserver that you have installed ? (asadmin version --verbose=true). If you are not using SJSAS8.1 UR2 p11, try to install that and see if you can still reproduce the problem. (It might be fixed in the latest patch)
If that does not work, then try the below workaround and see if this helps.
Start the appserver from Korn Shell (ksh), logout and check if the appserver processes are still there.
Cheers,
Vasanth
We're getting the same. Is this related to bug 6389609? (http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6389609). We're using the JVM parameter -XX:+PrintGCTimeStamps and rotate logs on a daily basis. These logs are then gzipped up so they are no longer available to the app server. The appserver nodeagent (which has that JVM parameter) coredumps every now and then. There are no error messages in the logs, it just coredumps.Are you using log rotation on your server.log?