Problem with DR via SYMCLI different behaviour on DMX2000 than DMX3

[dmaster]

Hello All,

We have a problem here with Disaster Recovery. All hour VMware ESX 3.02 servers see the R1 and the R2 of a SRDF-pair storage the same time.

So we can use DR and Vmotion in the same configuration.

- There is a Disaster on the symmetrix

- On the service console we do a manual failover for the Tier1 LUN with SYMCLI.

- Within ESX Server we change the advanced settings for the ESX server. "LVM.DisallowSnapshotLun from 1 to 0"

- Refresh you datastore

- Your datastore will be visible again with the same datastore name, but it uses now a different SCSI Target ID. (vmhba1:0:70 --> vmhba1:1:70)

- start virtual machines again

Above procedure works fine for EMC Symmetrix DMX 2000..

When we try the same procedure on a EMC Symmetrix DMX 3 system the SCSI Target ID will not change after a failover

Anybody familliar with this problem ? We have to get DR working on the DMX 3 system..

Regards,

Dennis Derks

[dmaster]

Hello All,

We found the solution!!

The DMX 3 storage gets confused when presenting the R1 and R2 at the same time to an ESX server.. it's actually the logical volume manager wich causing the problem.

The funny thing is that DMX 2000 does not get confused.. So now we are aware of this behaviour Disaster Recovery works fine.. without Diskresignaturing and without renaming your datastore or update your .vmx files.. just do a failover , reregister your VM's on the new ESX host in Virtual Center and power on the VM's

Regards,

Dennis Derks