Ok, customer has a Clariion without replication manager or
snapview.  We need to back up a couple of terabytes.  What I want
to do is mount the vxfs filesystems read-only on another Sun and backup
directly from that host via a couple of directly attached scsi LTO drives.

Can someone tell me why this will corrupt all our data, crash
every system in the machine room, and bring about the end of Western
Civilisation?  Or might it actually work?

[ Post a follow-up to this message ]

On 18 Lug, 01:13, w...@panix.com (the wharf rat) wrote:
>         Ok, customer has a Clariion without replication manager or
> snapview.  We need to back up a couple of terabytes.  What I want
> to do is mount the vxfs filesystems read-only on another Sun and backup
> directly from that host via a couple of directly attached scsi LTO drives.
>
>         Can someone tell me why this will corrupt all our data, crash
> every system in the machine room, and bring about the end of Western
> Civilisation?  Or might it actually work?

Since your FS will be mounted on another system, which will locally
cache the data, and you manage to activate the volume group, you'll
probably finish up with a corrupted FSs, changed on the go w/o
consistency from the original node. Even a raw device backup won't
work.  Probably the best thing is snap-shot and then locally or via
NFS remote mount the snap-shotted image.  Alternativeli Symantec/
Veritas sells a cluster solution which will give access to more then
on node in read/write mode, but isn't free.

[ Post a follow-up to this message ]