Oracle Database Discussions

Hi guys,

On my CentOS 6.6 system with Oracle 12c, a system that takes VERY limited load (used for dev and testing), when I start the instance, the taken physical memory is around 3.3GB. Then it starts fluctuating between 3.3GB and 7.5GB, for no apparent reason. This fluctuation takes please even when noone is using it at all.  Then overnight it stabilizes close to the limit, 7.9GB and stays there. I can't understand why it does that. Running "top" every 5sec (and comparing the 3.3GB state to the 7.8GB state) shows all top-memory processes at roughly the same occupancy, I can't see any process that has increased its footprint. The system runs Oracle and Tomcat appserver, nothing else. Tomcat is also completely stable.

top - 10:37:00 up 1 day, 11 min,  3 users,  load average: 0.07, 0.07, 0.04

Tasks: 197 total,   1 running, 196 sleeping,   0 stopped,   0 zombie

Cpu(s):  6.7%us,  2.8%sy,  0.0%ni, 90.2%id,  0.4%wa,  0.0%hi,  0.0%si,  0.0%st

Mem:  8001.801M total, 7837.074M used,  164.727M free,   25.270M buffers

Swap: 1999.996M total,  211.957M used, 1788.039M free, 1087.617M cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

1996 tomcat   20   0 4142m 1.1g 7184 S  0.0 14.1  17:26.92 java

4441 oracle    20   0 1372m 255m 252m S  0.0  3.2   0:10.83 ora_dbw0_oradb2

13891 oracle    20   0 1389m 232m 225m S  0.0  2.9  53:04.55 oracle_13891_or

21487 oracle    20   0 1368m 145m 140m S  0.0  1.8   0:05.80 oracle_21487_or

21847 oracle    20   0 1368m 143m 139m S  0.0  1.8   0:11.38 oracle_21847_or

4455 oracle    20   0 1380m 137m 121m S  0.0  1.7   0:40.79 ora_mmon_oradb2

4429 oracle    20   0 1364m 123m 123m S  0.0  1.5   0:03.78 ora_mman_oradb2

21829 oracle    20   0 1369m 123m 118m S  0.0  1.5   0:05.38 oracle_21829_or

4492 oracle    20   0 1375m 104m  94m S  0.0  1.3   0:22.76 ora_cjq0_oradb2

21844 oracle    20   0 1368m 100m  96m S  0.0  1.3   0:06.33 oracle_21844_or

I have the following settings:

AUTO memory management

kernel.small = 16TB (default value, I don't know why it's so large)

RAM=8GB

$ df /dev/shm

Filesystem     1K-blocks   Used Available Use% Mounted on

tmpfs            7340032 943088   6396944  13% /dev/shm

memory_max_target                    big integer 4G

memory_target                        big integer 1536M

The activity is so low that a query on v$resource_limit gives consistently something like this:

------------------------ ------------------- --------------- -------------- -------------- ----------

For SGA (v$sgastat):

------------- -------- --------

For PGA (v$pgastat):

aggregate PGA target parameter                 536

aggregate PGA auto target                      435

global memory bound                            100

total PGA inuse                                 54

total PGA allocated                             93

maximum PGA allocated                          359

total freeable PGA memory                       27

PGA memory freed back to OS                   8432

total PGA used for auto workareas                0

maximum PGA used for auto workareas             11

total PGA used for manual workareas              0

maximum PGA used for manual workareas            1

bytes processed                              24435

extra bytes read/written                         0

Please help me reach a sane selection of memory settings for this system.  I'm new to Oracle and I have to make do without formal training, regrettably.

Thanks a lot,

George