1.Firstly, how much “physical” memory is allocated:
$ /usr/sbin/prtconf System Configuration: Oracle Corporation sun4v Memory size: 16384 Megabytes ....
2. Is revolutions of the clock hand non-zero and increasing (say 1+ a minute). Revolutions of the clock is set to 0 on system startup. If it is non-zero but everything appears fine now then it indicates there may have been a problem in the past.
vmstat -s
....
843 revolutions of the clock hand
....
3. Scan rate in vmstat, consistently non-zero value in the hundreds+
vmstat 1 1000 | awk '{printf("sr = %d\n",$12);}'
4.Is current memory free less than a few hundred MB? pagesize = 8192, used for calculation below (sr in vmstat reported in pages).
$ /usr/bin/pagesize 8192(in GB):
$ sar -r 1 1000 | awk '{printf("memfree %.2f\n",($2*8192)/1024/1024/1024);}'
...
memfree 0.31
memfree 0.30
memfree 0.28
...
5. Is the ‘w’ (wait) column in vmstat consistently non-zero? w column:
$ vmstat 1 kthr memory page disk faults cpu r b w swap free re mf pi po fr de sr vc vc vc vc in sy cs us sy id 0 0 105 20514416 6943200 379 532 328 702 1539 0 26970 35 26 31 2 5342 14534 8187 3 4 93 1 0 93 30418504 12141008 33 32 0 0 0 0 0 0 0 0 0 1212 3252 1163 0 1 99 0 0 93 30418312 12141072 30 14 0 0 0 0 0 0 0 1 0 1172 1657 1089 0 1 99 1 0 93 30410120 12136976 30 14 0 0 0 0 0 0 0 0 0 1149 1618 1080 0 0 99
If yes to all above, what is using the most memory?
prstat -s size -Z
...
9705 grid 487M 470M sleep 59 0 0:00:00 0.0% oracle/1
9695 grid 487M 472M sleep 59 0 0:00:00 0.0% oracle/1
ZONEID NPROC SWAP RSS MEMORY TIME CPU ZONE
0 182 13G 13G 84% 6:07:50 6.2% global
...
