Zone memory issue

UNIX Solaris
1

We have a zone configured in our X4600 machine with memory capped to 16GB. Most of the time zone is running with high physical memory utilization. It seems from "top" command shows that the command "kernel" is locks 15GB phy. memory and not using swap memory. Whenever we restart the application running on this zone which releases the memory and in next 24hrs we are facing the same issue.

lab102sbx# top
load averages:  1.39,  1.31,  1.35;                    up 28+00:35:02                                                                                          06:29:17
101 processes: 100 sleeping, 1 on cpu
CPU states: 89.4% idle,  2.1% user,  8.6% kernel,  0.0% iowait,  0.0% swap
Memory: 16G phys mem, 2422M free mem, 278G total swap, 259G free swap

   PID USERNAME LWP PRI NICE  SIZE   RES STATE    TIME    CPU COMMAND
  8403 root       1  59    0 3684K 1892K cpu/9    0:00  0.01% top
  7043 sbxadm    19  59    0   26M 2280K sleep    0:14  0.00% igsmux_mt
 10922 noaccess  19  59    0  185M   45M sleep   44:57  0.00% java
  7044 sbxadm     9  59    0   54M   21M sleep    0:07  0.00% igspw_mt
  3580 root       2  59    0 5732K 2512K sleep   17.4H  0.00% cdm
 10172 sdb        3  59    0   54M   17M sleep    0:23  0.00% dbmsrv
  7045 sbxadm     9  59    0   54M   19M sleep    0:07  0.00% igspw_mt
  4603 sdb      114  59    0   26G   15G sleep   46:55  0.00% kernel
 10504 root       1  59    0 6632K 1972K sleep    3:20  0.00% controller
 10924 root       2  59    0 5192K  948K sleep    5:00  0.00% spooler
 10057 root      17  59    0   15M 5632K sleep    1:01  0.00% svc.startd
 11070 root       1  59    0 5400K 1340K sleep   10:44  0.00% processes
 10155 root      35  59    0   10M 4500K sleep    4:40  0.00% nscd
 11218 root       2  59    0 5028K 1168K sleep    2:05  0.00% logmon
 10931 root       1  59    0 5252K  660K sleep    1:29  0.00% hdb

bash-3.00$ /usr/sbin/zonecfg -z labsbx info
zonename: labsbx
zonepath: /export/zones/labsbx
brand: native
autoboot: true
bootargs:
pool:
limitpriv:
scheduling-class:
ip-type: shared
net:
        address: 10.61.184.120
        physical: e1000g0
        defrouter not specified
net:
        address: 10.61.184.121
        physical: e1000g0
        defrouter not specified
capped-memory:
        physical: 16G
2

It is using swap. try the

swap -l

command in the zone.

There are 2 GB of free memory in the zone from what you posted. It also does not look like there is a problem. Is rcapd running constantly? - what is the output of

rcapstat -z 5 3

(run from the global zone.)

That will tell us if you are actually "bumping heads" with memory limit.

Are worried that too much memory is being used? I don't see a big problem at all.

3

Jim,

Though the server have 2GB free memory it is only ~13% of the capped memory (16GB). So we keep gettinng high memory alert all the time.

bash-3.00$ rcapstat -z 5 3
id zone nproc vm rss cap at avgat pg avgpg
35 lab102sbx 99 35G 3429M 16G 257G 0K 22G 0K
35 lab102sbx - 35G 3429M 16G 0K 0K 0K 0K
35 lab102sbx - 35G 3429M 16G 0K 0K 0K 0K