Configure NAS Failover for Clusters of NFS Shares
When you configure NFS shares in a cluster for NAS failover, the master node manages StorNext NAS services. In the event of NAS failover, these services along with the NAS VIP are transferred to the new active master node.
By enabling NAS failover on NFS shares within NAS clusters, you are also requiring that all clients connecting to the NAS cluster are running NFSv4. This version of NFS is required to support lock recovery after failover.
In an environment that does not use NAS failover for NFS shares within NAS clusters, you can use both NFSv3 and NFSv4. See Manage NFS Client Versions.
Note: NAS failover is supported only on the Xcellis WFD (CentOS7) for NFSv4 clients. For environments exporting NFS shares on Xcellis WFD or G300 gateways, or on Artico or MDC servers, clients connect to the shares through the master node's static IP address.
Make sure to perform the following steps before configuring NFS shares in a cluster for NAS failover.
- Verify that all clients accessing the NFS shares are unmounted.
- Verify that all clients connecting to the cluster are running NFSv4.
If your clients are running on a Windows OS, you will need to download a third-party NFSv4 client for Windows.
- In addition to running NFSv4, verify that all clients are running CentOS versions 6.4, 6.5, 6.6, or 6.7. Only these versions have been validated in supporting lock recovery.
- Verify that all nodes within the NAS cluster are running on an Xcellis WFD (CentOS7). Only Xcellis WFD supports the NFSv4.1 lock recovery feature.
- Set a NAS VIP for the NAS cluster. Clients must connect to the NAS cluster using the NAS VIP. See Set the VIP for a NAS Cluster .
Note: To configure SMB shares for NAS failover, simply set a NAS VIP for the NAS cluster. See Set the VIP for a NAS Cluster .
For Xcellis WFD, StorNext NAS should be configured on the 10 GbE network port. In addition, the IP addresses reserved for both nodes of the NAS cluster and the NAS VIP must be configured on the 10 GbE network port, which is typically the LAN client network. If you use another network port, the NAS cluster will fail to configure correctly.
- Unmount all clients.
After configuring the NAS cluster for NAS failover, you must remount the clients through the NAS VIP.
- Log in to the controller console from the master node. See Access the Appliance Controller Console.
- At the prompt, enter the following:
nascluster set nfs-ha <yes|no> [force]
The parameter is:
yesto enable NAS failover on the NAS cluster.
noto disable NAS failover on the NAS cluster.
If this option is included, then the command is issued without confirmation
If this option is not included, then you must confirm the command.
Duplicate Request Cache
When NAS failover occurs in a NAS cluster of NFS shares, the duplicate request cache is not preserved and the new server cannot recover it. The cache is stored in memory.