Problems with volume manager disk group in a failover cluster (W2k8)

[kvk]

I have a 5,4 Tb ibm ds4100 array which is divided into 3 logical disks due to controller's 2 tb LUN limitation. I need to use the whole size in a failover cluster, and decided to use SFW. So far I've encountered several problems:

1. If Windows MPIO is installed and I install DMP DSM, then Veritas Storage Agent does not launch and no disks in the SFW console.

2. If I uninstall DMP DSM (so, the regular W2k8 MPIO drivers are left), then any attempt to create a cluster dynamic group results in a BSOD (error in msdsm.sys driver)

3. If I remove MPIO, then all disks on all paths can be seen in the console. I can create a cluster dynamic group, but then I encounter numerous failures trying to move this group from on cluster node to another and as a result the cluster "Volume Manager Disk Group" can't be brought online (I have 4 paths). I cannot connect the nodes directly to the storage, since I have another arrays in our SAN.

Has anyone had problems like these? If so, please describe your solution. Thanks in advance.

[James_V]

Please review Notes 17 & 24 from the Hardware Compatibility List in reference to your Array.
Also what version of Storage Foundation do you have installed? The Compatibility list above is for SFW 5.1

-James

[Jay_Kim]

1 & 2. Multiple different vendor multipathing solution cannot be used to manage the same device.
So only install VxDMP DSM if you want to use Veritas multipathing. (Refer to HCL for compatibility) (No need for MS MPIO separately)

3. Need to know what "numerous failures" refers to when deporting and importing the CDG (Cluster diskgroup) from 1 node to another to be able to figure out what's happening.