Normally that indicates that a disk has hardware problems.
You should look in /var/adm/messages
to look for io operation errors
and do a
iostat -En | grep -i soft
to look for hard errors.
Assuming it is a failed disk you will need to replade it with this procedure
http://sunsolve.sun.com/search/document.do?assetkey=1-9-73132-1
But if it is a spurious error then you can force it back to clean state with a
metareplace -e mirror submirror command