Relinking a Chunk to a Device After a Disk Failure

On UNIX, if the disk on which the actual mirror file or raw device is located goes down, you can relink the chunk to a file or raw device on a different disk. This action allows you to recover the mirror chunk before the disk that failed is brought back online. Typical UNIX commands that you can use for relinking are shown in the following examples.

The original setup consists of a primary root chunk and a mirror root chunk, which are linked to the actual raw disk devices, as follows:

% ln -lg 
lrwxrwxrwx 1 informix 10 May 3 13:38 /dev/root@->/dev/rxy0h
lrwxrwxrwx 1 informix 10 May 3 13:40 /dev/mirror_root@->/dev/rsd2b

Assume that the disk on which the raw device /dev/rsd2b resides has gone down. You can use the rm command to remove the corresponding symbolic link, as follows:

% rm /dev/mirror_root

Now you can relink the mirror chunk pathname to a raw disk device, on a disk that is running, and proceed to recover the chunk, as follows:

% ln -s /dev/rab0a /dev/mirror_root