- Make sure your have the right patches for the OS.
- Try both 1.3, and 1.4 JVM's. In the past, it took forever to HP to deliver
and stable 1.2, 1.3 JVM. (forever is relative to our project consistently
barfing in the lab)
- Does using only interpreted mode do anything? ( -Xint )
- Did you check the JVM release notes from HP? They have recommended kernel
tuning parameters.
When there is a hung state, try using glance or gpm to look at your java
process. Make a note of number of threads, the state of the threads, memory
usage, how the threads are blocked , number of files and sockets open as well
as socket state. It may not solve the problem, but it may point you in the
right direction.
Also - try turning OFF DNS on your box. Make sure you are not trying to
resolve hostnames furing each request. The underlying C function
gethostbyname() from the quotation I heard is not thread scalable.
-Tim
Allen Byington wrote:
Tomcat freezes after an indeterminate amount of time. It just stops responding to requests. Doesn't time out and the process doesn't stop. I've done everything that I can think of to stabilize this and was hoping someone had had a similar problem with stability under hp-ux. I've altered the amount of memory that java is starting with up to 256MB and there was no change. When the process hangs, it is truly hosed and can't be killed with anything other than a kill -9. The kill -3 to force a coredump doesn't seem to do anything at all and it stops writing to the logs long before I get to kill it so there isn't an error message of any kind. I'm desperate and looking for ideas. Any help that anyone could give would be greatly appreciated.
Thanks
Allen Byington
[EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]