SAP HANA Memory Usage – What should I really monitor?

With so many memory areas of SAP HANA, What should I really Monitor with respect to Memory usage of HANA database?

Below picture is key in understanding the memory allocation sequence.

hanamemory

You can control the behaviour of total memory allocated to SAP HANA database by using allocation limit parameter (global.ini -> memorymanager -> global_allocation_limit)

Unless you have additional software running on the HANA server or the HANA licensed memory is less than you physical memory, global_allocation_limit need not be set.

Default value is 0 and most of the times this is not set assuming that the server will be only used by HANA database. This means HANA DB can use the complete memory of the server.

globallmt

Memory allocation in HANA Database implements a pool concept. That is, memory is pre-allocated (reserved) from the operating system for performance reasons. By default, the HANA database will allocate up to approximately 90% of the available physical memory.

This will be used for queries that will need additional space at once. This will be RESERVED and not released back to Operating System.

Hence the reserved memory does not indicate any memory issues with the database.

Key figure to be monitored is the active memory usage of the HANA database:

memused

Same has been clarified in below SAP Note:

2081473 – HANA Resident Memory : High Memory Usage

You can monitor the current usage of memory by each tenant by using below query:

SELECT * FROM “SYS_DATABASES”.”M_SERVICE_MEMORY”

tenmem

Total memory usage of the whole database can also be monitored with HANA_MEMORY_OVRVIEW (SAP Note 1969700 ).

memoverview

You can also monitor the history of total memory usage using _SYS_STATISTICS, like example below:

memusedgg

To check if the SAP HANA Memory parameters are set optimally, you can use HANA_Configuration_Parameters_1.00.70+ (SAP Note 1969700).

Important SAP Notes:

1969700 – SQL Statement Collection for SAP HANA

1840954 – Alerts related to HANA memory consumption

2175606 – HANA: How to set allocation limit for tenant databases

2081473 – HANA Resident Memory : High Memory Usage

1999997 – FAQ: SAP HANA Memory

1998599 – How-To: Analyzing high SAP HANA Memory Consumption due to Translation Tables

2222276 – How-To: Activating SAP HANA Memory Fencing

1898317 – How to handle HANA Alert 1: ‘Host physical memory usage’

1813020 – How to generate a runtime dump on SAP HANA

SAP HANA Cockpit 2.0 – Installation and configuration

SAP HANA Cockpit is a Web-Based  tool for administration, monitoring and maintenance of SAP HANA systems.

Below we discuss installation steps and some key points with respect to cockpit:

  1. Download the required SAP file from SAP Market Place.cockpitdown
  2. Extract the SAR file and trigger installation using hdblcm.
    Note that installer installs a HANA instance (Express edition) of its own.
    More details hereco1
  3. Verify the details, enter the password and continue. Like any HANA installation user SYSTEM is created on the DB which will be used to login to Cockpit also.
    Along with SYSTEM, the other key user created is COCKPIT_ADMIN.
    co2.png
    Note that 51021 and 51023 are the default ports for Cockpit and Cockpit manager respectively.
  4. First it is required to login to Cockpit manager to assign yourself required roles using user COCKPIT_ADMIN.
    co3
    CO4
  5. Now you are ready to login to the HANA Cockpit.
    co5
  6. If you want to add additional users or resources (HANA systems) to cockpit, you can do so via SAP HANA Cockpit manager.
    co6
    co7

 

Important SAP Notes:

2496215 – You’re not authorized to open the SAP HANA Cockpit

2618617 – Launching SAP HANA Cockpit 2.5.5 in Firefox

2496215 – You’re not authorized to open the SAP HANA Cockpit

2185556 – FAQ: SAP HANA Cockpit

2595496 – Register a Resource in HANA Cockpit

2149157 – Unable to access the SAP HANA Cockpit due to error “Failure – Unable to load groups”.

 

R3trans and hdbuserstore

When you change the password of connecting user of SAP HANA database, same should be updated in the key file of the HDBUSERSTORE on the Application Server.

You should be logged in with <SIDADM> for this procedure.  Take care of this specially if you application server and database are running on same server.

#> hdbuserstore SET DEFAULT <hostname>:3<instance Number>15 <user> <password>

Note: User option -i along with the command if you want to enter the options (mainly password) interactively.

Example:

hdbuserstore SET DEFAULT hanahost:30015 SYSTEM Hana@123

Once updated check the list using the below command.

#> hdbuserstore list

DATA FILE : /home/appadm/.hdb/hanahost/SSFS_HDB.DAT
KEY FILE : /home/appadm/.hdb/hanahost/SSFS_HDB.KEY

KEY DEFAULT
ENV : HANAHOST:30015
USER: SYSTEM

You can now check the connectivity with the DATABASE using R3trans -d.

If R3trans -d reports error, check the connectivity with hdbsql using the key created above.

#> hdbsql -U <KEY>      #Key is “DEFAULT” in our example.

This should connect to the database without any issues.

You might receive error sql code  rcSQL = -10709 in case of connectivity issues. Please check the host name and port again in case of issues.

In case of any issues/questions please leave a comment below. Happy learning.

 

Important Notes:

2250144 – FAQ: SAP HANA Secure User Store

2563762 – hdbuserstore still using old hostname after updating

2399996 – How-To: Configuring automatic SAP HANA Cleanup with SAP HANACleaner

2653777 – Phyton connection to HANA database using hdbuserstore-key fails

2416706 – Check file permissions for SAP HANA client hdbuserstore on application server

2130496 – Application does not start after takeover-HANA Disaster Recovery

 

2709350 – How To set a HANA System License via HDBSQL