cancel
Showing results for 
Search instead for 
Did you mean: 

Intermittent Initial Load Slowness for Random T-Codes Affecting Random Users

06-25-2025 8:29 PM
1561 views 6 comments Go to solution
0 Likes
SAP Managed Tags
Subscribe

Dear SAP BASIS/HANA Experts,

In our S/4 HANA production environment, we are encountering intermittent initial load slowness for various transaction codes (T-codes), affecting different users on a random basis. The issue occurs when a user executes a T-code for the first time during the day. Subsequent executions generally respond promptly. However, the issue may reoccur for the same or different users with different T-codes on the following days.

Our observations and findings are as follows:

  • The behavior is unpredictable, affecting random users and random T-codes.

  • The slowness is temporary, typically lasting 30–40 seconds, occasionally up to 1–2 minutes.

  • By the time the issue is reported, it has usually resolved, making it difficult to activate traces or collect dumps.

  • Analysis of STAD data reveals very high DB request times (sequential reads) during slowness against the affected user.

  • Affected users appear to be stuck in the “DYNPLOAD” action, as observed in SM50.

  • Program, screen buffer, table buffer parameters value are sufficient in ST02, No SWAP.

System Environment:

  • Standard architecture: S/4 application and HANA database hosted on the same server

  • Total RAM: 2 TB

  • HANA Memory Allocation Limit: 1600 GB

  • SAP Application Memory: 180 GB

  • Recent upgrade from 1.5 TB to 2 TB RAM completed prior to issue onset

We would appreciate your inputs and recommendations on how best to diagnose and address this behavior, particularly given the inconsistency and unpredictable nature of the symptoms.

Looking forward to your expert guidance.

 

0 Likes

Accepted Solutions (1)

Accepted Solutions (1)

SAPBeginner1234
Explorer
0 Likes

The issue is resolved now after executing SGEN.

Answers (1)

Answers (1)

mamartins
Active Contributor
0 Likes

The SAP HANA DB have very strict specifications regarding the ratio RAM/CPU, depending of the INTEL CPU architecture. The new RAM amount still respects that ratio?

Also, there are several situations that increasing the RAM modules count will lead to decreasing the RAM speed. Could you check that?

The new RAM modules are evenly distributed for all sockets, respecting the NUMA architecture?