High memory utilization on two system when running database

[itrc55]

Hi,

I facing problem with high memory utilization on two system (mph015,mph016) , which working as veritas cluster server (VCS)

When we run the database on one server or both the memory used percentage going up to 99% from 36%.

 BL870c i2 "HPUX 11.31

On System A

#hastatus -summ

-- SYSTEM STATE
-- System               State                Frozen

A  system A               RUNNING              0
A  system B              RUNNING              0

-- GROUP STATE
-- Group           System               Probed     AutoDisabled    State

B  cvm            system A               Y          N               ONLINE
B  cvm             system  B               Y          N               ONLINE
B  misprodSG       system A               Y          N               ONLINE
B  misprodSG       system B               Y          N               PARTIAL

Now on system B the memory utilization is 36% because the database state on this system is PARTIAL

I've already modified some kernel parameters like

filecache_max                3250710937  10%         Imm (auto disabled)
filecache_min                1625354240  Default     Auto

shmmax                       1073741824  1073741824  Immed 

To

shmmax                       8589934592  8589934592  Immed

but the problem is still, please find the attached file for some information (kctune, kmeminfo, ....)
How i can resolve that ?

Thanks,

(Attachment removed due to being HP Restricted or Confidential)

[Dennis Handly]

Why do you think you have a problem?

You have 21.6g for users and 8.5g for the kernel.

[itrc55]

Hi,

Why do you think you have a problem?
Because the system it's rebooted every 2 days.

????????

[Dennis Handly]

>Because the system it's rebooted every 2 days.

What causes you to reboot the system?

[V. Nyga]

Hi,

you have running 37 oracle processes - is this wanted?
If yes, then maybe it's time to buy more RAM.

With vmstat you can check (at least in UX 11.11) if your server is paging out data to swap (listed under 'po').

If yes - then you need more RAM (or less processes).

V..

P.S. What's about the cpu usage?

[itrc55]

Hi,

No CPU utilization

#vmstat 5 10
procs memory page faults cpu
r b w avm free re at pi po fr de sr in sy cs us sy id
2 0 0 4103122 30074 5 0 16 16 4 0 363 5098 32673 2641 2 1 97
2 0 0 3991229 29863 14 0 17 13 0 0 388 4992 30207 2652 2 1 97
2 0 0 3991229 30092 19 0 27 65 17 0 4776 4987 42848 2523 2 1 97
2 0 0 3847627 32169 34 0 40 20 4 0 1807 4904 34706 2589 2 1 97
2 0 0 3847627 32267 13 0 20 7 0 0 654 4789 30891 2406 1 1 97
6 0 0 3994725 32177 7 0 5 1 0 0 213 4765 28526 2492 1 1 98
6 0 0 3994725 32167 2 0 4 0 0 0 68 4762 31373 2495 2 1 97
1 0 0 3514738 32137 0 0 0 0 0 0 21 4794 29020 2460 1 1 98
1 0 0 3514738 29781 30 0 1 0 0 0 5 4757 25089 2355 2 1 98
1 0 0 3504548 29719 16 0 17 124 46 0 6975 4808 25868 2409 2 1 97

#swapinfo -tam
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 16384 0 16384 0% 0 - 1 /dev/vg00/lvol2
dev 16384 1337 15047 8% 0 - 0 /dev/vg00/lvol19
reserve - 1435 -1435
memory 31001 30053 948 97%
total 63769 32825 30944 51% - 0 -

Thanks

[itrc55]

Hi,

What causes you to reboot the system?

it's rebooted without user intervention because this the VCS behavior to save the database on one node and failover to second node.

And Don’t asked me anything about the VCS because I don’t have any experience.

Thanks

[Dennis Handly]

>the VCS behavior

As in Verilog Computer Simulator?  Are there any logs that explain the reason for the reboot?