1. The previous networking mode is that two Windows servers form an MSCS cluster and provide email exchange service. The back-end is an H3C storage device connected to the servers by using the iSCSI ports.
2. After all services are stopped, map the H3C storage device and the S5300 to a VIS storage device. The three storage devices are used as three LUNs whose capacities are 500 MB, 500 MB, and 1.4 TB. The H3C storage device is used as source mirror LUN. The data on it is migrated to another two LUNs.
3. After the migration is complete, the VIS storage device is shut down. The S5300 is mapped to the Windows server and the S5300 is connected to the server through multiple paths. The server still employs the MSCS cluster mode and provides email exchange service.
4. After one server in the cluster is started and the UltraPath software is installed, the Windows 2k3 server can identify three LUNs of 500 MB, 500 MB, and 1.4 TB. (The LUN IDs are Q, S, and T. Q indicates the quorum disk.) However, only the two file systems of 500 MB LUNs can be identified, while the file system of the 1.4 TB LUN cannot be identified. The 1.4 TB LUN cannot be used.
Step 1 Uninstall the UltraPath software and connect the S5300 to the server by using a single optical fiber. The fault persists.
Step 2 Map the S5300 to another server and find that the LUNs can be identified. It proves that after data migration, the data on the S5300 is available.
Step 3 Map the S5300 to the cluster server. Start the cluster after the quorum disk (disk Q) is identified. Import the LUN resources into the cluster, and then the file systems of all LUNs can be identified.
Step 4 Start two cluster servers after the LUNs are identified. Switch the two servers and test the other server. Customer’s services are normal.
The MSCS cluster service does not have LUN resources.