CCI Command Device Best Practice

To minimize contention and to maximize availability so that disaster recovery or data replication operations are not delayed or denied, the following are the "best practices" for configuring the Command Device (CMDDEV).

 

• There should be two or more CMDDEV.
  • CCI will use an alternate command device when there is an error from the raw IO to the primary device.
  • Older DF subsystems such as AMS1000 only have 2 CMDDEV - so all users have to share them. The AMS2000 series has up to 128 CMDDEV.
  • RAID subsystems like USP, USP-V and VSP can have many CMDDEV - so each Application/Host should have its own CMDDEV(s).
  • There is a command that can be used to alternate to the next command device in the list of command devices. The command is: horcctl -C.
  • You can display the current active CMDDEV using the command horcctl -D.
• The CMDDEV should be a LUN that is not "used" by any host. It must not be formatted with a file system nor mounted.
• The CMDDEV should be setup as Raw devices.
• If TC/A is being used, there should also be multiple command devices on the remote controller.
• For disaster recovery applications, the multiple command devices should have at least two independent paths.
• There should be at least two physical ports.
• The physical ports should have independent network paths.
• The command devices can be a LVI/LUN as small as 36MB.
• Under Windows, always use \\.\CMD syntax. For more details, refer to Hitachi Command Control Interface User and Reference Guide, MK-90RD7010-04.