The following best practices ensure a robust Storage Foundation Cluster File System High Availability (SFCFS HA) or Storage Foundation (SF) for Oracle RAC campus cluster:
Tag all the mirrored volumes in the campus cluster with appropriate site names. VxVM allocates storage from the correct site when creating or resizing a volume and when changing a volume's layout if the volume is tagged with site name.
All volumes that have data required by the application must be evenly mirrored. Each site must have at least one mirror of all volumes hosting application data, including the FlashSnap log volume.
In Flexible Storage Sharing (FSS) environments, if possible, use multiple mirrors at each site since storage availability is tied to host availability due to local DAS or internal disks being used. For more details on FSS, see the Storage Foundation Cluster File System Administrator's Guide.
Do not enable site consistency on VxVM snapshot volumes.
Use redundant links for storage and private interconnects. DWDM can be used for storage and heartbeat together. Another redundant DWDM link can be used to prevent single point of failure. Separate switches and multiplexer / de-multiplexer devices should be used.
Use Coordination Point Server as the third coordination point.
Use the procedure for online replacement of coordination points, to replace disk based or Coordination Point Server based coordination points.