I'm having trouble trying to figure out why the DNS process' memory is so high. We upgraded to Solaris 10 not too long ago on both our internal/external DNS servers. Both internal/external DNS configurations are the same. Both running the same version of BIND, named.conf files are the same, same server models, memory, HDD, etc. All matching. The difference is, two different Unix admins did the upgrades. One no longer works with us, who setup the external DNS and the other who installed 10 on the internal DNS is no help. He insists he configured it correctly. Our external servers are not experiencing this problem. The DNS process never goes over 60mb. The internal servers process either dies at 2.6gb or until swap space runs out. When that happens, the server needs a reboot. I'm not a Solaris expert and I need some assistance with troubleshooting. I'm guessing it has to do with shared memory. Anyone ever run into something similar?

Internal Server

load averages: 0.09, 0.06, 0.05; up 245+05:26:25 09:40:54
61 processes: 60 sleeping, 1 on cpu
CPU states: 99.1% idle, 0.2% user, 0.7% kernel, 0.0% iowait, 0.0% swap
Memory: 2048M phys mem, 37M free mem, 4001M total swap, 2436M free swap

PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND

10596 root 5 59 0 1973M 696M sleep 61:49 0.13% named

External Server

load averages: 0.02, 0.02, 0.01; up 224+05:35:54 09:42:09
57 processes: 56 sleeping, 1 on cpu
CPU states: 98.3% idle, 0.9% user, 0.8% kernel, 0.0% iowait, 0.0% swap
Memory: 2048M phys mem, 75M free mem, 3076M total swap, 3074M free swap

PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
521 root 5 59 0 53M 44M sleep 30.1H 0.57% named