Tuesday, December 7, 2010

Clustering in SAP Systems

Steps for High Availability Failover Clustering for SAP Systems
1)Configure the Internal Disks RAID configuration
2)Fill in general worksheet with values like IP Addresses (7 IP addresses, 2Private Network, 2Public Network, 3Virtual)
3)Install OS (Windows Server 2003) on Node A
4)Install OS (Windows Server 2003) on Node B
5)Modify BOOT.INI on NodeB (Change the timeout =5 in the boot loader file so that there is no simultaneous contetion of shared drives)
6)Configure disks directories on RAID1
7)Storage Manager configuration (Fibre channel configuration)
8)MSCS pre install verification
9)configure the server service change memory parameter to zero
10)Set the page file size to 5 times the RAM size
11)Modify the BOOT.INI for 4GB tuning NOT NEEDED
12)Remove drivers and protocols Remove unnecessary drivers or disable them (Spooler), only one protocol TCP/IP is needed,
13)Install MSCS on NodeA at end of installation reboot NodeA
14)Install MSCS on NodeB at end of installation reboot NodeB
15)Installation of troubleshooting tools MOM
16)Windows NT service pack installation
17)DNS and transport directoy configuration On the transport host create the directory \usr\sap\trans If no instance will be installed on the transport host share this directory as SAPMNT
18)Installation of IE, MMC, page etc
19)Install the latest SAP Kernal patches
20)Apply the post SP4 DLLs Latest Windows cluster patches
21)Correction of Network bindings Test the host files i.e move the groups from one Node to the other. Ping response should come from the public network, Chnage the order in which the cards are listed as:1)Public Network, 2)Backbone Network, 3)Private Network
22)MSCS verification test the service using the CASTEST utility, power down one of the nodes and check the logs.
23)Create the installation user account
24)Installing SAP and the DBMS
*Preliminary Work Filling in all the details from worksheets.
*SQL Server Install- Install SQL Server on NodeA Start SQL Server on NodeA -Register ServerA- Perform simple queries- Stop SQL Server
* Install service pack on NodeA
*Install R/3 setup tool From SAP kernal start the program Sapinst.exe
*Install SAP R/3 Central Instance and Database Instance give necesary parameters, This will take several hours, Reboot NodeA
* Run the cluster Failover wizard on NodeA Start=>Programs=>SQL Server=>Failover Cluster Wizard. Here give SA Password, SQL Service Account, IP Address (VIRTUAL IP FOR ACCESSING THE DATABASE ON THE PUBLIC NETWORK), Server Name (ITSSQL)
After installation reboot NodeA and NodeB
When system restarts log on to NodeA as Installation user (sapinst)
Move all resources back to NodeA
Configure two shares SAPLOC and SAPMNT pointing to \USR\SAP
Restart the windows services: SAPOSCOL and SAP
Restart the databse service in the cluster administrator
All the resources for SQL Server are now displayed in the cluster Administrator anc can be moved between Nodes.
*SQL Server cluster conversion
Install the Cluster Conversion tool on NodeA, provide SAPSYSTEMNAME ITS, \users\adm\install, it will ask do you want to logoff yes
Now Install cluster conversion tool on NodeB
SAP Cluster conversion:
Creates the SAP cluster group
■ Copies tools to the SAPCluster directory
■ Sets the SAPService to manual
■ Makes sure that the SAPOsCol service is started from the SAPCluster directory