VM problems with secondary datastore

[officeboy]

I have a server that uses local storage for primary disks, and an NFS share for backup locations. Now this NFS target isn't anything special (netgear NV+), and it gets shutdown, moved, cables fall out etc.
Before being virtualized that wasn't a problem, since windows thinks of the backup targets as more or less a removable drive. But now the VM freaks and the whole thing goes offline. The VM stops responding, and won't come back up till the NAS is restored. This is a very undesirable dependency.
Is there a way to break this that anyone knows of, make ESX know that this disk is not critical to operations.

 

[NetJunkie]

Mount the NAS from inside the VM, not as a datastore with a disk. Then Windows will reconnect as needed else it's like walking up to a server and pulling out a direct attached SCSI disk. Bad things happen.

 

[lopoetve]

APD is not supported, so what NetJunkie said We try to pause all IO to storage is restored - if possible.

 

[officeboy]

Looks like I have trouble either way then, since SBS 2008 will not schedule a backup to a network drive/share.

At least I know now that it's called All Paths Down (ADP)
Thanks

 

[geiger]

I'm not too familiar with SBS or the Netgear, but can you utilize iscsi volumes within the VM.

 

[officeboy]

Server 2008 and SBS 2008 backup do not work well with networked drives. iscsi or NFS are treated more or less the same in windows.

 

[Chilly]

Server 2008 and SBS 2008 backup do not work well with networked drives. iscsi or NFS are treated more or less the same in windows.

iSCSI is not(or rather should not be, but I doubt SBS is any different) treated as a network drive in Windows, but again I don't know if your NAS supports it.

Once set up, Windows should see it as a block level device.