SCM 4.0 and livecache sharing the same server ?

[Former Member]

Hello,

We currently have our SCM 4.0 system on a seperate server from our livecache 7.4.3 build 41. Both run on solaris servers. We are getting a new server that would be big enough to run both on this new server. I know some shops have SCM and livecache running together on the same server. I am looking for any feedback, good and bad on the option of running them together.

Thank you,

Steve

[roland_mallmann]

Hi Steve,

well, ofcourse having just a single machine would be nice from for example the OS side, just a single machine to update.

But when choosing a single machine to hanlde both loads in parallel, the machine should at least be twice in size (CPUs, RAM) and scaleable for future needs.

What kind of hardware are you talking about? Can you give us an overview of your current machines and the new one?

Are we talking production system(s) here?

Do you have any additional Application servers in the same landscape?

Regards,

Roland

[Former Member]

Hi Roland,

Thank you for replying.

This is our production system that we are talking about.

Our current livecache server is a Sun V1280 with 12x U3/900 Mhz CPUs and 96GB Ram.

Our current SCM4.0 DB and CI system is running on Sun E6800 with 24 CPUs and 128 GB Ram. We also have 9 instances running on 3 app servers. The app servers all run on SUN 12x U3/900 Mhz CPUs with 96GB Ram.

The new machine is a SUN SF6800 1.2 GHZ with 24 CPU Sparc 3 single lore with 192 GB Ram.

Thanks,

Steve

[roland_mallmann]

Hi again,

thanks for the update! I mentioned that the new machine should at least have the combined power of the old machines, but ofcourse this is not set in stone and depends on the current useage levels of the different machines.

How many CPUs/RAM is still available on the current liveCache during heavy loads? Same question for the SCM DB server?

I take it the Appl. Servers would not be moved to the single machine?

Any plans to add more load to the system in the near future (adding plants or anything similar)?

Regards,

Roland

[Former Member]

Hi Steve,

just to add to my colleague Roland's reply: you can find out whether this single server will be sufficient for your needs, by the QuickSizer tool. This tool helps in hardware sizing for your business requirements.

SAP customers can find it at: https://service.sap.com/quicksizer

Thanks,

Ashwath

[Former Member]

Hi Roland,

Our livecache server has crashed in the past during heavy loads.

I usually notice that the Data Cache occupancy % will get to 100% and the history/undo occupancy will be increased as well.

In the past during our heavy loads we were constantly getting CPU alarms in our SCM system which were taking all our resources. We have never had any issues with RAM. A couple months back we reduced the number batch jobs running at the same time. Our functional team would run all their SNP or DP jobs at once. We now stagger them to run 50 or 60 at a time. This has helped and we no longer have the CPU alarm issues.

You are correct about the app servers, these will not be moving.

I am not aware of any future growth to the system at this time.

Thanks,

Steve

[Former Member]

Thank you Ashwath for your post.

Steve

[roland_mallmann]

Hey Steve,

It might be an idea to have SAP take a look at your liveCache config to see what kind of performance improvements are still possible using that machine.

Is liveCache configured to use all available CPUs currently? And did it use all of them during heavy loads?

Are the CPUs on the SCM system the same as on the new system (i.e. same speed etc.)?

---

Just a question on the side: what is the reason you're getting the new machine? Is it due to the (occupancy/performance issues you experienced with your current liveCache machine?

Regards,

Roland

[former_member229109]

Hello Steve,

You wrote:

"Our livecache server has crashed in the past during heavy loads.

I usually notice that the Data Cache occupancy % will get to 100% and the history/undo occupancy will be increased as well."

Hints to solve the reported to problems::

1. What do you mean about the liveCache server crash?

Did you have the hardware problems? Or liveCache was crashed?

Did you create the OSS ticket for the analysis of the liveCache crash?

Please see the SAP Note No. 354700 to find what information we need for

the analysis of the liveCache crash.

It will be helpful at first to check the Stack backtrace information

from the knldiag.err or corresponding knldiag file, if the liveCache was

crashed.

2. If you noticed that the Data Cache occupancy % will get to 100% and the

history/undo occupancy will be increased as well, please see the SAP

note 631629, and please check if you have too "old" OMS versions?

And what application is responsible for that?

What is the value of the liveCache parameter LOAD_BALANCING_CHK?

Please see the SAP note No. 695721.

Did you see the long running LCA routines?

What is MAXCPU & OMS_HEAP_COUNT liveCache parameter's values?

< You wrote, that your functional team would run all their SNP or DP

jobs at once. ?? We need to check the application scenario

& variants, data selections of those jobs. What those jobs are

doing? Did you create the ticket to the component 'SV-BO-SCM'

on this issue? >

Did you check the internal < TA //om17 >& external < //ccr > inconsistencies on your system?

Please review the SAP note no. 425825.

Did you run the programs /sapapo/ts_lcm_reorg,/sapapo/ts_lcm_reorg_snp and /sapapo/ts_lcm_cons_check_all to check the consistency of time series data?

For this problem it will be helpful to logon to your system & check the status. Please create the OSS ticket to the component 'BC-DB-LCA' or 'BC-DB-LVC'.

***

For SAP liveCache documentation in English < See the SAP note 767598 >:

http://help.sap.com/saphelp_nw04/helpdata/en/f2/0271f49770f0498d32844fc0283645/frameset.htm

Did you use the DB Analyzer to find the bottleneck in liveCache?

***

Also the question for you is:

-> Did you create this thread to solve the 1. & 2. problems?

Thank you and best regards, Natalia Khlopina

Message was edited by:

Natalia Khlopina

[Former Member]

Hi Roland,

We have had SAP dialed in previously after we had our performance issues as well as when we crashed last year. We had sent them the knldiag error files as well. They did have us change some parameter values and suggested our functional team run their jobs differently. The crash I am talking about was the last crash we've had and that was one year ago.

The reason for the new box is we have almost reached our max capacity on the current box. I did not mean to open a can of worms here, just trying to get some feedback on the pro's and con's of having the two systems share one server.

Thanks,

Steve

[roland_mallmann]

Hi Steve,

thanks again for the background information.

I just wanted to make sure that you make sure that you take a good look at your current heavy loads on the liveCache and SCM system, so that you can make an educated guess as to whether your new machine would be able to handle both.

I will ask a couple of colleagues if they know if certain bigger liveCache customers have such a single machine SCM & liveCache setup.

Regards,

Roland

[Former Member]

Hi Roland,

I appreciate your help on this question and your advice.

Thanks,

Steve