How to find resident memory via sap solution manager System Monitoring

[fmet]

Hello all,

I'm looking for the Alert(s) which are used for resident memory size in as well HANADB as MAXDB. The purpose for my call is using SAP Solution Manager - System Monitoring.

I have found the following text "Real memory size refers to resident memory size (or resident set size). This is the actual amount of physical RAM being used by the process".

However I can find the corresponding Alerts in the corresponding templates provide by SAP Solution Manager > Application Operations > System Monitoring.

I have seen the term Resident memory is used in Focused Insights, so my assumption is there should some information available in SAP Solution Manager. (working with SAP Solution Manager 7.2 SPS09).

Best regards

Frans Metselaar

[timkarl89 · Accepted solution]

Hello Frans,

I re-checked this. I found the spot in Focused Insights:

In my case I've got a total available memory of 1.000GB; currently used by the HDB are 259 GB.

The resident memory is about 880 GB. The concept is explained in detail in this document:

"Resident memory is the physical memory actually in operational use by a process. Over time, the operating system may “swap out” some of a process’ resident memory, according to a least- recently-used algorithm, to make room for other code or data. Thus, a process’ resident memory size may fluctuate independently of its virtual memory size. In a properly sized SAP HANA appliance there is enough physical memory, and thus swapping is disabled and should not be observed"

This graphic might be helpful:

There is no metric for the resident memory in the templates default. You could create a custom metric based on the SQL statement:

Select HOST, round((USED_PHYSICAL_MEMORY + FREE_PHYSICAL_MEMORY)/(1024*1024*1024), 2) as "Physical Memory GB", round(USED_PHYSICAL_MEMORY/(1024*1024*1024), 2) as "Resident GB" from PUBLIC.M_HOST_RESOURCE_UTILIZATION

...or you are looking for an existing HDB alert ID. I would prefer the second option.

Back to your question:

Since I found no suitable default HANA alert for this I would use the "host physical memory usage". If you are facing an out-of-memory situation, you are looking for the causing process right? I would not care if this is an SAP HANA or non-SAP process -> your database is affected.

A guide for creating custom metrics can be found here (the procedure is equal in 7.1 and 7.2). Next to this I would always take an existing metric as reference: Go to templates -> switch to expert mode -> Data Collection.

E.g. something like Memconsumption of the name server is just collecting an existing HDB alert:

Kind regards

Tim

[SriKrishna]

Memory Usage of Services HANA Metric tracks the actual memory used by indexserver, not resident memory at OS level.

you can use the OS level metric, which i shared above.

[SriKrishna]

hi,

use host metric - File system utilization including File System Cache %

This the amount of physical memory free on the host in percentage of total memory available. It also includes the memory currently used for the file system cache, which will be released by the OS if necessary.

[fmet]

Hi,

Thank you for your answer.

I will certainly make use of your answer.

Best Regards,

Frans

[timkarl89]

Hello Frans,

I found the following standard metrics which might be helpful. My systems- and contents basis is 720 0011:

• Host Physical Mem Usage (of course this contains OS and other processes next to HDB)
• Memory Usage of HANA services

If I understood you right, you are looking for something like total memory consumption by the SAP HANA system excluding server processes right?

AS far as I see you need to define a custom metric to collect this information.

Kind regards

Tim

[fmet]

Hello Tim,

I also found that metrics, but it is quite hard for understanding all the different flavors.

But, yes , I'm looking for the actual memory used by the processes. And not only for HANA but also for MAXDB.

Do you have an example for such a custom metric. Should be very helpful.

Best regards,

Frans