Hello everyone -
We have a database server which is choosing to swap though there is plenty of
free memory.
Any pointers on what to check are much appreciated..
Environment:
Red Hat Enterprise Linux Server release 6.7
Oracle EE 12.1.0.2
This server has 2 instances
This server is dedicated for Oracle workload, 2 database instances are running
from this server.
Server has 512GB memory available, 38% (192GB) is used by Oracle.
Here is how the memory allocation is:
Instance#1
SGA: 97G
PGA: 63G
Instance#2
SGA: 97G
PGA: 63G
AWR confirms the instance memory usage is about 40% of the server's physical
memory. No other process is running on the server other than the monitoring
tools like Splunk/OEM agent, which is not using a lot of memory.
Most of the memory is locked up in OS cache.
Though there is free memory from time to time, we see the swap being used..
$ more /etc/sysctl.conf |grep kernel.shm
kernel.shmmax = 338723604480
kernel.shmall = 4294967296
kernel.shmmni = 4096
OEM confirms the memory utilization is about 40% (or less) for as long as we
see it.
Trying to figure what might trigger the server to start using the swap while
there is plenty of free memory.
Appreciate your insights..
Thanks in advance
-Upendra
