How to expand (Scale Up) SAP HANA system
This lesson describes how you can expand your HANA system.
After completing this lesson you will be able to understand the basics of scale out and configuration of a distributed system.
You can find detailed information in scale out at:
http://help.sap.com/hana_appliance
SAP HANA scale-out architecture is used to deal with larger amounts of data or for higher availability.
If you need more memory or more CPU power beyond the limitation of a single physical hardware box you can use a distributed landscape consisting of multiple hosts.
A host is the server or blade on which you create an individual node of a system.
SAP HANA studio allows you to monitor all hosts in a distributed landscape.
Image courtesy: help.sap.com
During installation, directories for data (default /usr/sap/<SID>/global/hdb/data) and log area (default /usr/sap/<SID>/global/hdb/log) are defined.
The next directory level is mnt00001, mnt00002 and so on, where each worker host uses exactly one directory.
Installing Master and Slave Servers of multiple storages mounted at mnt00…, because the number of directories does not change when new services are added.
The next level is the actual volume hdb00001, hdb00002, with one directory per service.
Note: There is no storage volume assigned to the Standby server.
Parameter System Default Value
Lowest free instance number available on the Instance number host /usr/sap
Note: You can only accept this default value during the installation if you install a single-host system.
You must change this default path if you plan to create a distributed system.
Installation path
Home directory /usr/sap/<SID>/home
Number with value x+1where x is the highest ID of the user existing user ID on
the current host Group ID 79
Home of the login: shell /bin/sh
Path for storing the data volumes: <inst_path>/<sid>/global/hdb/data
Path for storing the log volumes: <inst_path>/<sid>/global/hdb/log
Image courtesy: help.sap.com
To rename the hosts of a distributed system, you need to perform the rename on each host of the distributed system.
When setting up a distributed system, you need to make sure that the servers are synchronized to the same time. To achieve this, you can set up a Network Time Protocol (NTP) service.
For testing and debugging its possible to copy a scale out landscape to a single node.
Some helpful SAP notes on scale out: 1681092( Multiple SAP HANA databases on one appliance), 1661202 (Support for multiple applications on SAP HANA), 1666670 (BW on SAP HANA – landscape deployment planning)
