[SunHELP] Disk Suite - How to change a failed drive

Steve Sandau ssandau at gwi.net
Wed Aug 26 17:14:51 CDT 2009

Sandwich Maker wrote:
> " From: Steve Sandau <ssandau at gwi.net>
> " 
> " []
> " 
> " I replaced three disks today in an E3000 and was able to run metastat 
> " and look at the output to determine the bad disks. (The metadevices on 
> " the bad disk show up with a "metareplace" command in the output.) I 
> " physically replaced the bad disks and ran "devfsadm -C" to tell the OS 
> " to find the new disks. Then I ran the metareplace command as shown in 
> " the output of metastat with the addition of the "-e" flag in the 
> " command. (That flag replaces the bad device with a new device in the 
> " same position.) You should find that in the man page as well if I am not 
> " being clear enough.
> if a bad disk also has a metadb copy on it, you have to handle that
> separately.

Ah, yes, that is true. Running 'metadb' before all the above steps and 
using metadb -d <slice> to remove any metadbs showing as 'unknown' ought 
to take care of that. Then you would recreate them afterwards.

The other step that needs to be done before the metareplace is to copy 
the partition table from the good mirror to the new disk with 'prtvtoc 
/dev/dsk/c0t0d0s2 | fmthard -s - /dev/rdsk/c0t1d0s2' where t0 is the 
good disk and t1 is the new one.


More information about the SunHELP mailing list