This section describes the prerequisites for switchover of operating or standby status of storage.
The following disk resources are the targets of switchover.
Dynamic LUN mirroring
The replication is automatically configured.
LUN prepared in the storage unit
LUN replication settings need to have been made beforehand between the operational storage and backup storage.
The LUN replication ratio between operating and standby storage states must be 1:1.
If the operating disk resource is a disk (TPV or FTV for ETERNUS) with Thin Provisioning attributes set, the standby disk resource must also be a disk with Thin Provisioning attributes set. Similarly, if the operating disk resource is a disk with thick provisioning attributes set, the standby disk resource must also be a disk with thick provisioning attributes set.
Operating disk resources must be connected to physical L-Servers.
Disk resources that are not registered in storage pools or are not connected to L-Servers are not processed.
The switchover of disk resources is processed according to the replication relationship described in the replication definition file created in advance. Disk resources that are not described in the replication definition file are not processed.
If LUNs are added or the storage configuration is modified, it is necessary to edit the replication definition file.
Standby disk resources must be detected by Resource Orchestrator. If LUNs can be accessed from the server, Resource Orchestrator cannot detect them. Do not register detected disk resources in storage pools.
The storage unit identifier to enter in the replication definition file (IP address for ETERNUS, NetApp, or EMC CLARiiON, or SymmID for EMC Symmetrix DMX Storage or EMC Symmetrix VMAX storage) must not be of the same configuration.
In this case, storage units with the same IP address or SymmID as an operating storage unit cannot be used as standby storage units.
For configurations with NetApp storage units using the MetroCluster function for storage replication, switchover cannot be performed with this function.
By switching the operating and standby storage using the MetroCluster function, operation of physical L-Servers can be restored.
To access operating and standby storage units from servers with physical L-Servers running on them, it is necessary to set the fibre channel switch in advance.
If the storage unit is ETERNUS, no settings are required in advance.
The time required for switchover is relative to the number of L-Servers using operating storage units and the number of disk resources being used by L-Servers.
It is recommended that a test be performed in advance to confirm the time for restoration from storage unit failure.
