Hello all,
We have a T5 sparc machine running Solaris 10 branded non-global zones. The global zone is running 11.1. We are running a an application that we suspect could be causing a memory leak in one of the zones. Eventually, memory is used up and swap space is completely full. When this occurs the entire T5 sparc server has to be rebooted. Memory resource capping has been implemented to prevent a full machine reboot.
My question is what tools can we use in the Solaris 10 zone to troubleshoot or find memory leaks? As of now the zone is doing just fine with no evidence of extensive memory use. The /var/adm/messages log file would display the following messages when the issue occurs.
Oct 2 01:55:29 spechost2 genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to grow stack for pid 4877 (alertagent.8.0.2)
Oct 2 01:55:46 spechost2 tmpfs: [ID 518458 kern.warning] WARNING: /zones/vkaytoo2/root/etc/svc/volatile: File system full, swap space limit exceeded
Mdb cannot be used because the device files used for this utility exists on the global zone. Do we need to use tools in the global zone to detect shared memory use?
Any ideas or links to resources would be greatly appreciated. BTW. I have been searching the web on this subject and are good for global-zones.
-Mike/Ismael