i normally do something like this, which might be less work, and doesn't
require an intermediate copy of the file system (unless you'd like
the backup, which you might).  i was going to send this privately but
i'll post it to the list in case i've slipped up somewhere.

add the new discs to the file server, leaving the original one as well,
being sure to use distinct SCSI IDs.

create a new file system on the new devices, initialise it and set allow mode

	filsys newfs [w2w3]	# or whatever your configuration is for the new devices
	ream newfs			# be very careful to ream the right name!
	allow

now you've got access to both the original main and empty newfs
file systems at the same time from the cpu server or terminals.

copy from one to another, preserving permissions and ownership

	# mount existing file system on /n/boot so as to avoid copying files in /net and /dev(!)
	9fs boot

	# mount the destination file system on /n/tapefs (or some other name)
	mount -c /srv/boot /n/tapefs	newfs		# note the `newfs' argument to select other than `main'

you can do some test file creations, writes and removals on /n/tapefs.

	# copy the files, preserving permissions
	disk/mkfs -a -s /n/boot /lib/proto/allproto | disk/mkext -u -d /n/tapefs

in this last command, again be sure that /n/boot looks right and /n/tapefs was correctly
mounted from the file server, and that you've typed it in correctly
(and that what i gave you makes sense),
because you're in allow mode without a safety net.

then `halt' the file server, remove the old disc and enter config mode.
i'll use the same scsi IDs as above but with the old device out you could
renumber with appropriate changes.

	config w2		# or whatever is appropriate
	# re-enter your ip ipmask ipgw and service name
	filsys main [w2w3]
	end