Host outages during SVC Codeupgrade: Missing Warning for misconfiguration and missing SVC check – command

All backend-paths of each mdisk needs to have the same Host-LUN-ID (SCSI-ID). In case of different IDs on different paths of the same mdisk, SVC does accept this without a warning. But during a SVC Code-Upgrade, this leads to a offline-SVC-pool-situation over e.g. 5 minutes. This results: no host io, filesystem-errors, host-reboots, fsck, data-restores or dataloss.

Of course, the SVC customer has to take care for correct configuration of the connected storage systems. Nevertheless SVC should offer methods to discover this “time bombs”.
By the way, other Products (e.g. ESX or Veritas DMP) run without issues with different LUN IDs on same LUNs.

Request for Enhancement:
1) SVC should either …
- run and update without issues with once accepted backend-/connection-configurations
or
- not accept configurations that leads to outages during updates (mdisk IDs)
or
- at least warn if there is a misconfiguration in the backend-connection.

2) SVC should offer the SVC administrator a command to view and check all mdisk LUN-IDs (on each path). The output of lscontroller or lsmdisk is not sufficient.
The SVC / IBM Support already told me, that the mdisk Host-LUN-ID from the backend can be found in a support-snap- or generated-dump-file, but is only visible to IBM stuff. Therefore the information is available inside the SVC-Cluster. Please make it visible to the SVC administrator.