I have completed several LUN transitions from 7 Mode to Cluster Mode using the 7 Mode Transition Tool successfully, but this past weekend we had a failure. The Windows cluster hosting the LUNs that I was trying to migrate is the only one that already has a few LUNs on cdot, as well as 7-mode LUNs I'm trying to transition. The 7MTT choked when it saw that the IQN already existed, so it didn't map the 7-mode LUNs. I mapped them manually before restarting the Windows hosts, but then we had LUN ID conflicts between the 7-mode and cdot LUNs. I had to rollback to transition.
I believe I have a resolution to this problem, but would love to get feedback from any of you that might have a comment. Here's what I'm planning to do:
Have the DBA’s turn off and disable SQL services for the instance hosted on cdot LUNs.
Use SnapDrive to disconnect the cdot LUNs.
Unmap the cdot LUNs and remove the cdot initiator groups, which of course removes the LUN IDs.
Have the Server Engineering team shut down all nodes of the Windows cluster.
Allow the 7-Mode Transition tool to transition the 7-Mode LUNs as it normally does, including LUN mappings (as a backup, I would have commands ready to manually do this if necessary, but this time with the LUN IDs added to the commands).
Reboot the cluster nodes and verify that the 7-Mode LUNs transitioned successfully and are functioning.
Reconnect the LUNs that had already been on cdot via SnapDrive, which should assign them new non-conflicting LUN IDs.
Anyone see a gap or flaw in this plan? Any recommendations?
Thanks. So if I remove the cdot LUNs and the initiator groups and then remap them after the other LUNs transition, are you indicating that won't resolve my issue? We'll still potentially have a conflict due to the signature? If so, how can I strip the existing signature off of the cdot LUNs?