Switchover Failure of Dual FC Switches

Publication Date:  2012-07-22 Views:  102 Downloads:  0
Issue Description

Under the environment with dual switches, working nodes are configured to check hard disks every one minute. One of the nodes cannot be started properly when the other node is switching over its services, causing the switchover to fail.

Product and version information:
  • S5000 series (with dual controllers and FC host ports)
The networking of the environment with dual switches is as shown in Figure 1.
Figure 1 Networking of the environment with dual switches

Alarm Information

In the ISM, choose Fault > Fault Management, and a large number of Host port link down alarm messages can be found under Fault List.

Handling Process
  1. Replace the FC port module of the storage device to see whether the fault persists.
  2. If the switchover of dual switches succeeds, the fault is removed. Otherwise, replace the FC cables between the AS and storage devices to see whether the fault persists.
  3. If the switchover of dual switches succeeds, the fault is removed. Otherwise, replace FC switches.

     

     NOTE:

    For the steps on replacing FC switches, see corresponding manuals supplied with the product.

     

Root Cause
  1. It is found that the link indicator of FC host ports light on and off frequently in an interval of one minute.
  2. In the ISM, choose Performance > Bit Error Rate Statistics and bit errors on FC host ports can be found.

Conclusion:

  • FC host ports linkup and linkdown frequently due to faulty FC host ports, FC cables, or FC switches, causing LUNs to switch over continually. When the AS performs "pvscan", timeout occurs for the read/write operation due to frequent switchover of LUNs, causing the dual switches fail to switch over.
Suggestions

It is recommended to set domains for FC switches to avoid this problem.

END