swap space reduced drastically

[Former Member]

Hi Techies

We have solaies 10,oracle10g ,BI

Our production server is working fine but suddenly the swap space will reduced from 259GB to 100GB in 15 min of time span( when i checked in df -h) and then apllication will hang.

We have a work around like we will restart the application then the swap space will be released to the allocated

In this due time if i check prststat (process status) IGS process is consuming 150GB of space only.

what could be the reason for reducing the swap space drastically???

Is this the problem of application or Operating system

I have checked the note (724713)

Thank You

Haroon

[Former Member]

Thanks you

everybody this issue has been resolved

we have upgraded the kernel patch level and igs patch level

[Former Member]

This something New for me too.

I heard these type of problem due to improper os patching or kernel parameters specific to Solaris.

Since you have opened the message with OSS.Please share the solution with us.

When are you upgrading IGS.

Good Luck..

Amit

[Former Member]

Hello Haroon

Please refer the OSS note

Note 124199 - ST06 display - Paged in [Kb/h] / Paged out [Kb/h]

And I believe this is related to un Solaris issue.You may need to go back to Sun through SAP problem ticketing.

Hope this help your query .

Good Luck

Amit

[Former Member]

Thanks for the reply Amit /Mark

I have raised the OSS message they have replies to upgrade the IGS 7.0 13 patch

But my question is

We have allocated 269 Gb of swap space. The used space and available space do not change rather the allocated space itself is reducing.

What could be the reason for this.

[markus_doehr2]

Solaris uses an "early swap" algorithm, means, it will allocate the swap early before the physical memory is used completely - so what you see is normal.

Why so much space is used after that time is an application issue.

How many users use the system, how much physical memory do you have and what are the profile parameters you configured (em/initial_size_MB, ztta/roll_area_dia/nondia)?

Markus

[Former Member]

Thanks for the quick reply

df -h ( detail)

Filesystem size used avail capacity Mounted on

swap 258G 394M 258G 1% /tmp

swap 258G 56K 258G 1% /var/run

swap 258G 0K 258G 0% /dev/vx/dmp

swap 258G 0K 258G 0% /dev/vx/rdmp

From the above the allocated size swap is reduced drastically when the issues arise.

There are 57 users connected to the server

the physical memory is 65 GB

( i also checked the note 425207 profile parameter

em/initial_size_MB =8192

ztta/roll_extension 2000683008 Byte

[markus_doehr2]

You mean the "available space" is reduced? And you said: IGS is using 150 GB RAM?

Markus

[Former Member]

Yes the avaliable swap is reducing

[markus_doehr2]

Ok - then the swap is used.

And you IGS is really consuming 150 GB (Gigabytes) of space?

Markus

[Former Member]

yes

IGS(internet graphic server) is consuming the 150 GB of swap space

[Former Member]

IGS server is slowly consuming the swap space to maximum

My question is how the allocated swap can be reduced.

example we have given swap space of 258GB this is allocated

This swap space is reduced not the avaliable

and the only process which is consuming the reducing allocated swap is IGS

Hope you understand me

[markus_doehr2]

I can tell you: this is not normal.

I would upgrade the IGS to the latest available version and see if the problem persists. If it does, I would open an OSS call (BC-FES-IGS). It's not common using THAT much memory to render graphics.

Markus