Greetings,

I have an E250, running Solaris 8 and disksuite, fully populated with 6
identical internal SCSI drives. The first two are set up in a RAID 1 meta
device, the next three are set up in a RAID 5 meta device, and the last
drive is set up as a hot spare for the RAID 5 meta device.

A few days ago, my SMART monitoring system started warning that one of the
drives (c0t11d0s0) in the RAID 5 array is about to fail and I'm seeing
many read errors in the logs.

So far though, this hasn't been enough to convince disksuite to swap it
out with the hot spare; metastat says that the drive is fine (I suspect
this is a combo of retryable read errors, a logging filesystem and not
much loading at the moment).

I'd like to remove the failing drive from the RAID 5 array before it
completely fails, but I can find nothing in the disksuite docs that allow
me to force fail it (mark it as "bad" and compel disksuite to swap in the
hotswap drive in it's place--something easy to do in any linux distro).

I can find nothing in the documentation about, this, but is it possible to
force the issue with disksuite by metareplacing the failing drive with the
hot swap drive?

I.e, I'm thinking something like:

metareplace d40 c0t11d0s6 c0t12d0s6

Where d40 is the RAID 5 meta device, c0t11d0s6 is the failing drive, and
c0t12d0s6 is the hotspare drive associated with d40.

This is a live server so I'm trying to avoid downtime. If I can flip the
bad drive out of the RAID 5 array, I can then hot swap it with a new one
without having to reboot.

TIA for any and all pointers/suggestions/etc...