Planning for Smart Assist for SAP Netweaver
To use Smart Assist for SAP Netweaver, you must install and tune the AIX® operating system according to the SAP OSS Notes and the official SAP installation guides.
To view what versions of SAP Netweaver are supported, see the Support matrix for PowerHA SystemMirror Smart Assists topic.
SI Number | Single point of failure | Volume group requirement | File system mount point | IP address requirement |
---|---|---|---|---|
1 | SAP global file system | A minimum of 1 volume group for each, attached to both nodes. | /export/sapmnt/<SID> /export/usr/sap/trans |
One free IP |
2 | ABAP central services instance | A minimum of 1 volume group for each, attached to both nodes, or a local file system that is maintained by the user on both nodes. | /usr/sap/<SID>/ASCS<nn> | Installed with a dedicated virtual IP |
3 | Java™ central services instance | A minimum of 1 volume group for each, attached to both nodes, or a local file system that is maintained by the user on both nodes. | /usr/sap/<SID>/SCS<nn> | Installed with a dedicated virtual IP |
4 | ABAP ERS instance | A minimum of 1 volume group for each, attached to both nodes, or a local file system that is maintained by the user on both nodes. | /usr/sap/<SID>/ERS<nn> | Installed with a dedicated virtual IP |
5 | Java ERS instance | A minimum of 1 volume group for each, attached to both nodes, or a local file system that is maintained by the user on both nodes. | /usr/sap/<SID>/ERS<nn> | Installed with a dedicated virtual IP |
6 | Application server instance | A minimum of 1 volume group for each, attached to both nodes, or a local file system that is maintained by the user on both nodes. | /usr/sap/<SID>/DVEBMGS<nn> | Installed with a dedicated virtual IP |
- The volume group names that are displayed in Table 1 are just for demonstration, and you can replace them with a valid volume group name.
- To support older system installations ASCS and SCS instances can share a virtual IP. Also, both ERS instances can share a virtual IP. This setup is not recommended and applies only for Dual Stacks.
Before you can start Smart Assist for SAP Netweaver, review the following information:
- Ensure that your hardware infrastructure is set up for redundancy.
- When using shared disks, you must verify that all volume groups that become part of a resource group are attached to both nodes. You can manage local and shared volume groups by using C-SPOC.
- Configure and tune all nodes as described in the SAP installation guide of your release and version of SAP.
- Ensure that user IDs and group IDs of the SAP OS-users are the same on both nodes. You can manage users, groups, and password with C-SPOC (smit cspoc).
- Create the basic two node PowerHA SystemMirror cluster by using SMIT or the clmgr command.
- Install the SAP Instances on your primary node. All instances must be installed with their own virtual IP by using the ./sapinst SAPINST_USE_HOSTNAME=<IP alias> command (exceptions see Table 1).
- Prepare the second node. Ensure that the /usr/sap/sapservices file is the same on both nodes.
- Ensure that sapcpe program is configured such that all instances can operate independent from their availability of NFS during runtime.
- The service sapstartsrv of all instances must be running so that they can be discovered properly by the smart assist. This setting is the SAP default after the installation.
- The Sap Global directory must be accessible from all the nodes and setup to be highly available.
- The SAP Admin user <sid>adm must be the SAP default.
- Apply the SAP HA licenses to run SAP Netweaver instances on the failover nodes as required by SAP.
- Verify that SAP Netweaver instances
can run successfully on all the failover nodes by completing the following
steps:
- Set up the failover nodes for accessing the SAP global file system.
- Create SAP users on all failover nodes with same uid and gid values.
- Run the SAP Netweaver instances on failover nodes by manually varying on the volume group, which contain the file system for the instance directory (on a shared disk) and set up the logical IP addresses as an IP alias by using the ifconfig command.
- Remove the
MAIL
variable information from the .login file that is set by SAP. You must have SAP administrator authority to remove theMAIL
variable. - When you have mail there is additional output from commands that are run by
PowerHA SystemMirror Smart Assists that can cause the
parsing of the command output to fail. To fix this issue, add the following code at the start of the
the .cshrc
file:
setenv MAIL "/var/spool/mail/$LOGNAME" setenv MAILMSG "[YOU HAVE NEW MAIL]" if ( $?prompt == 0 ) then # This is a shell script .... do nothing else # This is interactive if ( -f "$MAIL" && ! -z "$MAIL") then echo "$MAILMSG" endif endif
Note: If you are not using the csh shell, then use the appropriate variable for the prompt. - You must verify that the .cshrc and .login files do not produce any output text.
- To perform SAP administrative functions, add the SAP admin user to the hacmp group by using the SMIT interface (smitty chuser) or the chuser command. This process allows the SAP admin user to access the PowerHA SystemMirror binary files and ODMs.
Before installing any SAP Netweaver instances, you must set up the global file systems on all the cluster nodes by using NFS cross mounts.
If you plan to also add the SAP Global directory and the SAP Transport directory to this cluster, the following items apply in addition:
- Set up the following mounts on a node where the volume group is
hosting file system that is varied on:
- /export/sapmnt/<SAP System ID>
- /export/usr/sap/trans
- Set up the following NSF cross mounts on all nodes, for example:
- mount <nfsip>:/export/sapmnt/<SAP System ID> /sapmnt/<SAP System ID>
- mount <nfsip>:/export/usr/sap/trans /usr/sap/trans
Note: Smart Assist for SAP cross mounts file systems related to SAP global file systems such as, /export/sapmnt/<SAP System ID> and /export/usr/sap/trans use NFS version 4 on all cluster nodes.NFS is used for the SAP global file system. If the /sapmnt directory is inaccessible due to NFS failure caused by a network or interface failure on a particular node, PowerHA SystemMirror tries to move any of the SAP instances running on that node to another node. Therefore, the SAP instances might fail to stop on that node, resulting in a cluster event script error. If this problem occurs, you must manually complete the fallover by using the SMIT utility Recover From PowerHA SystemMirror Script Failure.
- Verify that ABAP and Java central services instances as well as all ABAP and Java enqueue replication server instances do not run on the same node while using Smart Assist for SAP Netweaver.
- The ERS is cluster controlled (no polling) and is installed with a virtual IP.
- To configure a SAP instance run Smart Assist for SAP Netweaver on the node where that particular instance is running.
- Verify that SAP Netweaver 2004s
instances can run successfully on all the failover nodes by completing
the following steps:
- Set up the failover nodes for accessing SAP global file system.
- Create SAP users on all failover nodes with same uid and gid values
- Run the SAP Netweaver 2004s instances on failover nodes by manually varying on the volume group that contains file system for the instance directory (on a shared disk) and set up the logical IP addresses by using IP alias.
- Apply any required licenses to run different SAP Netweaver 2004s instances on the failover nodes.
Now the nodes are configured to run Smart Assist for SAP Netweaver on the node where that particular instance’s sapstartsrv process is running.