PRIMECLUSTER Wizard for NAS (PRIMEQUEST) 4.2 Configuration and Administration Guide - Linux -

Contents

Chapter 5 Notice

When using the PRIMECLUSTER Wizard for NAS, take notice of the following:

• If you set UnLockflag to "always" in scalable operation, the Unlock processing is executed during split brain. Then, NFS re-locking is executed, and the lock is set again, so this does not disrupt ongoing operations. But the NFS relocking might slow down operations.
• If you specify IP instead of an NAS host name in /etc/fstab.pcl, the NFS lock control might fail. You must specify a host name. The host name should be the same as the one that is registered in NAS.
• When you register a host name of a cluster node in the NAS hosts file, the host name should be the same as the one that is registered in that cluster node. If it is different, the lock control might fail.
• The PRIMECLUSTER Wizard for NAS temporarily creates a file in /tmp. Do not modify access privileges to /tmp or delete the temporary file in /tmp. You also need to reserve enough space in /tmp.
• When you use a clustered NAS, you need to specify an export volume in the "etc/exports" file, so takeover/giveback will be properly performed.
• When you use a clustered NAS, NAS does not temporarily respond during takeover/giveback. To prevent failover, you need to set both the ServerTimeout and Timeout of the RemoteFileSystems resource to about 300 seconds. The value varies depending on an environment.
• In the following cases, Netapp resource does not goes into Warning state.
  - When control facility for the NAS-Cluster is not used in standby operation.
  - When failover of the NAS device has been caused by manual operation or failure other than failure of network.
  - When failover of the NAS device has been caused by the "Negotiated Failover for Networking" function of DataONTAP7.2 or later.
  If NAS device after failover is recovered (giveback), discontinue monitoring Netapp resources so that failover doesn't occur again while recovering.

Contents