Overview of shared disk and tape devices
In a PowerHA® SystemMirror® cluster, shared disks are external disks connected to more than one cluster node that are used for application shared storage.
In a nonconcurrent configuration, only one node at a time owns the disks. If the owner node fails, the cluster node with the next highest priority in the resource group node list acquires ownership of the shared disks and restarts applications to restore critical services to clients. This ensures that the data stored on the disks remains accessible to client applications.
Typically, takeover occurs within 30 - 300 seconds. This range depends on the number and types of disks used, the number of volume groups, the file systems (whether shared or Network File System (NFS) cross-mounted), and the number of critical applications in the cluster configuration.
When planning the shared external disk for your cluster, the objective is to eliminate single points of failure in the disk storage subsystem. The following table lists the disk storage subsystem components, with suggested ways to eliminate them as single points of failure.
| Cluster object | Eliminated as single point of failure by... |
|---|---|
| Disk adapter | Using redundant disk adapters |
| Controller | Using redundant disk controllers |
| Disk | Using redundant hardware and LVM disk mirroring or RAID mirroring |
In this section, you perform the following planning tasks:
- Choosing a shared disk technology.
- Planning the installation of the shared disk storage. This includes:
- Determining the number of disks required to handle the projected storage capacity. You need multiple physical disks on which to put the mirrored logical volumes. Putting copies of a mirrored logical volume on the same physical device defeats the purpose of making copies. For more information about creating mirrored logical volumes, see Planning shared LVM components.
- Determining the number of disk adapters that each node will contain
to connect to the disks or disk subsystem.
Physical disks containing logical volume copies should be on separate adapters. If all logical volume copies are connected to a single adapter, the adapter is potentially a single point of failure. If the single adapter fails, PowerHA SystemMirror moves the volume group to an alternate node. Separate adapters prevent the need for this move.
- Understand the cabling requirements for each type of disk technology.
- Adding the selected disk configuration to the cluster diagram.
- Planning for configuring a direct fibre channel tape unit attachment as a cluster resource.