In this section we will set up a highly-available NFS mount target using the NetApp Cloud Volumes Service for Google Cloud, which is a cloud-native third-party service that integrates with Google Cloud to provide zonal high-availability for shared file systems. Complete the following steps to create a Cloud Volume, mount it on each cluster node, and prepare the shared file system for installation of the SAP instances.

  1. Follow the steps provided in the Quickstart for Cloud Volumes Service to purchase and enable the Cloud Volumes Service in your Google Cloud project.
  1. Once the Cloud Volumes Service has been successfully enabled, navigate to the Cloud Volumes → Volumes page and click Create at the top to create a new Cloud Volume.

  1. Enter the following parameters and click Save at the bottom of the page to create the Cloud Volume. The icon indicates that the default option is chosen.
Field
Value
Volume Name cv-sapnfs
Service Type CVS
Region <Deployment region> (e.g., us-east1)
Zone <Zone> (e.g., us-east1b)
Volume Path cv-sapnfs
Service Level Standard-SW
Volume Details 1024 GiB
Protocol Type NFSv3
Make snapshot directory visible Unchecked
Shared VPC configuration Unchecked
Use Custom Address Range Unchecked
Export Rule 1

Allow ‘Read & Write’ on 10.20.0.0/22

Note: We are exposing the NFS export to the entire lk-subnet subnetwork for simplicity, but the scope of the export rules can be made more fine-grained to only allow access from the cluster nodes which are mounting the NFS share.

Allow automatic snapshots Unchecked

The cv-sapnfs Cloud Volume will appear on the Cloud Volumes → Volumes page once it has been successfully provisioned.

  1. Note the Mount Target for the NFS export. This will be used to mount the shared file system on the cluster nodes. In the following steps we will use <Mount Target IP> to refer to the IP address of the NFS server listed here.

  1. Execute the following commands on both node-a and node-b to create the /sapnfs mount point for the cv-sapnfs NFS share. Add the given mount entry to /etc/fstab on both nodes (replacing <Mount Target IP> with the IP obtained from the Cloud Volumes page in the previous step) so that each cluster node mounts the shared file system at boot.
# mkdir /sapnfs
# echo “<Mount Target IP>:/cv-sapnfs	/sapnfs	nfs	nfsvers=3,proto=tcp,rw,sync,bg	0 0” >> /etc/fstab
# mount -a

Verify that the file system has been mounted successfully on both nodes.

# df -h | grep sapnfs
<Mount Target IP>:/cv-sapnfs  1.0T  1.2G 1023G   1% /sapnfs
  1. Execute the following command on node-a to create the subdirectories within /sapnfs which will contain the shared SAP files.
[root@node-a ~]# mkdir -p /sapnfs/{sapmnt/SPS,usr/sap/trans,usr/sap/SPS/{ASCS10,ERS20}}

Verify that the subdirectories were created successfully.

[root@node-a ~]# du -a /sapnfs
0       /sapnfs/sapmnt/SPS
0       /sapnfs/sapmnt
0       /sapnfs/usr/sap/trans
0       /sapnfs/usr/sap/SPS/ASCS10
0       /sapnfs/usr/sap/SPS/ERS20
0       /sapnfs/usr/sap/SPS
0       /sapnfs/usr/sap
0       /sapnfs/usr
0       /sapnfs
  1. Execute the following commands on both node-a and node-b to create soft links from the SAP file systems to their corresponding subdirectories under /sapnfs.
# mkdir -p /sapmnt /usr/sap/SPS
# ln -s /sapnfs/sapmnt/SPS /sapmnt/SPS
# ln -s /sapnfs/usr/sap/SPS/ASCS10 /usr/sap/SPS/ASCS10
# ln -s /sapnfs/usr/sap/SPS/ERS20 /usr/sap/SPS/ERS20

Verify that the soft links have been created successfully on each node.

# ls -l /sapmnt /usr/sap/SPS
/sapmnt:
total 0
lrwxrwxrwx 1 root root 01 Jan  1 00:00 SPS -> /sapnfs/sapmnt/SPS

/usr/sap/SPS:
total 0
lrwxrwxrwx 1 root root 01 Jan  1 00:00 ASCS10 -> /sapnfs/usr/sap/SPS/ASCS10
lrwxrwxrwx 1 root root 01 Jan  1 00:00 ERS20 -> /sapnfs/usr/sap/SPS/ERS20

The shared file systems are now prepared for installation of the SAP instances.

Feedback

Thanks for your feedback.

Post your comment on this topic.

Post Comment