Sun Microsystems 2.1 manual server# mfs repair -r -d /dev/rdsk/c1t4d0s2

In the output above, make note of the SCSI disk address (similar to the number c1t4d0s2). Chapter 3 of the Sun MediaCenter UltraSPARC Hardware Manual has instructions for mapping SCSI disk addresses to physical disks.

2.Open the appropriate Multipack enclosure, locate the failed disk, and remove and replace it.

Documentation for your disk subsystem tells you how to replace a disk. There is no need to halt the server or reduce its load while replacing the failed disk.

Caution –It is only at a point after you have received the “disk failed” message described above that you can safely remove and replace a failed disk. Do not confuse this definitive message with, for example, “soft errors,” such as SCSI read errors. If you remove a disk that is not failed while the server is in operation, you can cause extreme disruption to the server, possibly including a hang or a panic.

3.Following disk replacement, enter the following mfs repair command to restore data from the parity disk to the new disk:

server# mfs repair -r -d /dev/rdsk/c1t4d0s2 -v

The address c1t4d0s2 is an example, consistent with the preceding example commands.

4.When the preceding command completes, enter the following command to return the server to normal playback mode, phasing out parity mode.

server# mfs repair -p -d /dev/rdsk/c1t4d0s2 -v

Upon successful return of the preceding command, the server is in normal playback mode and you can, once again, withstand a single disk failure.

You can combine the -rand -poptions, shown in the previous two example commands. For example:

server# mfs repair -rp -d /dev/rdsk/c1t4d0s2 -v

With a command such as the preceding, you can combine repair and return-to- normal-playback in a single command.

Chapter 8 Administering MFS Disks 8-11