We are noticing some instances of NETSMART 2000 slowly growing native heap ( greater than 2 Gig ) while java vm itself seems fine and within -Xmx limits. One side effect of this problem manifests when other applications (unix or another vm) needs to launch a process or need to use swap and fails.

Another problem we see is, once we hit the issue, we are not able to run diagnostic tools as the swap space is maxed out.

[] > java -version
java version "1.5.0_06"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_06-b05)
Java HotSpot(TM) Client VM (build 1.5.0_06-b05, mixed mode)

I also have attached the output of 'pmap -x' 'jmap -info' 'jmap -heap' as run on the process.
After more investigation, the following seems true ( the bug manifests if the client is started on a V210/240 and doesnt manifest on Ultra60s).

gamma (U60 - server) : yemen ( V210 -client) = reproducible
gamma (U60 - server) : gamma ( U60 - client) = not reproducible
griphon4(V210-server):gamma ( U60 - client) = not reproducible
griphon4(V210-server):griphon4 ( V210 -client) = reproducible

Maybe V210/240s need more patches ?

I have attached 2 files ( U60.showrev.out and V210.showrev.out).

Could you please compare and tell us if we need more patches for our V210/240s ?