I/O data path protection to redundant
Important:
A multi-path driver is a driver, such as
AVT is a built-in feature of the RAID controller module
The storage management software has defined behavior for the following configurations, based on host type. For details on which operating systems support which type of failover, refer to your
Note: If you want to change the default AVT settings without changing the host type, contact Customer and Technical Support.
A pair of active RAID controller modules are located in a storage array. When you create a virtual disk, you assign a RAID controller module to own the virtual disk (called
The preferred RAID controller module normally receives the I/O requests to the virtual disk. If a problem along the data path (such as a component failure) causes an I/O to fail, the multi-path driver will issue the I/O to the alternate RAID controller module.
When AVT is enabled and used in conjunction with a host multi-path driver, it helps ensure an I/O data path is available for the storage array virtual disks. The AVT feature changes the ownership of the virtual disk receiving the I/O to the alternate RAID controller module.
Once the I/O data path problem is corrected, the preferred RAID controller module will automatically reestablish ownership of the virtual disk as soon as the multi-path driver detects the path is normal again.
When AVT is disabled, the I/O data path will still be protected as long as you use a multi-path driver. However, when an I/O request is sent to an individual virtual disk and a problem occurs along the data path to its preferred RAID controller module, all virtual disks on that RAID controller module will be transferred to the other RAID controller module instead of just that particular virtual disk.
Storage arrays in this scenario have no failover protection. A pair of active RAID controller modules might still be located in a storage array and each virtual disk on the storage array might be assigned a preferred owner. However, virtual disks are not allowed to move to the alternate RAID controller module.
When a component in the I/O path fails, such as a cable or the RAID controller module itself, I/O cannot get through to the storage array. The component failure must be corrected before I/O can resume. (Switching virtual disks to the alternate RAID controller module in the pair must be done manually.)
Hosts using operating systems without failover capability should be connected to the storage array so each
Changing RAID Controller Module Ownership of a Disk Group or Virtual Disk