cancel
Showing results for 
Search instead for 
Did you mean: 

HANA Memory Used Discrepancy

0 Kudos

Hello,

I was wondering if anyone knows how to reconcile the reported MEMORY USED on the OVERVIEW tab in the studio with the number reported by Memory Consumption of all row tables and Memory Consumption of all column tables transactions on the SYSTEM INFORMATION tab. The Overview tab reports a significantly higher number. I understand that there is always working memory and that's why one should only use 50% of available memory but the difference between memory consumed by all loaded tables and reported on the overview table is not 2X but 10X.

Accepted Solutions (0)

Answers (2)

Answers (2)

former_member182277
Contributor
0 Kudos

Hello,

Please try this also. I might helpful for you.

system -> system information tab -> double click on the Memory of all row tables/Memory of all column tables.

You will able to see the available/used memory.

Memory count can vary due to the compression rate.

Regards,neha

0 Kudos

Hi

I am not talking about compression factor at all or comparing it to ECC table sizes, if you look at my original post. Basically I am expecting:

Loaded memory of all row tables + loaded memory of all column tables = Memory used on the overview tab.

Even accounting for working memory space requirement,

Loaded memory of all row tables + loaded memory of all column tables = 50% of Memory used on the overview tab.

However I see Memory used = 10 X  (loaded memory of all row tables + loaded memory of all column tables)

former_member182277
Contributor
0 Kudos

Hello,

Could you please restart your database once and check if discrepancy is showing or not and let me know.

Many times some memory is reserve for the work that we have done.

Hope it is helpful.

Regards,Neha

lbreddemann
Active Contributor
0 Kudos

Kraft BI Team wrote:

for working memory space requirement,

Loaded memory of all row tables + loaded memory of all column tables = 50% of Memory used on the overview tab.

However I see Memory used = 10 X  (loaded memory of all row tables + loaded memory of all column tables)

So what?

HANA uses a memory pool management that takes it as granted to have all the hosts memory at avail all the time. Since allocating and especially deallocating memory on OS level is pretty expensive, HANAs memory allocation on OS level will typically grow over time and usually be at least double the size of the data on disk.

Looking at the memory used on OS level really isn't worth too much on HANA.

regards,

Lars

rama_shankar3
Active Contributor
0 Kudos

Lars: Interesting -  What is your SAP HANA hardware configuration (make of server, memory, no.of CPU's)? Depending on the hardware configuration, the compression algorithm is different. I have in my research material details on how HANA handles storage space on different hardware. Will look for it and get back to you.

Regards,

Rama

lbreddemann
Active Contributor
0 Kudos

Rama Shankar wrote:

Depending on the hardware configuration, the compression algorithm is different. I have in my research material details on how HANA handles storage space on different hardware. Will look for it and get back to you.

Hmm... that is something I haven't heard so far. Would be interesting to hear about that (although I have to say that I heavily doubt it...)

Let me know if you have some details on this.

best regards, Lars    

rama_shankar3
Active Contributor
0 Kudos

I think that you are confusing yourself with the data compression ration for column tables vs. space utilization. Please review the administration editor -> system -> storage data by data / log / trace and do the math. You will need admin privilidges to view this data within the HANA studio.

Hope it helps..

Rama