Hi,

I am trying to restore a full system, which I backup with the ufsdump command, in to a different unix box.
I partitioned that box(second one) exactly the same as the first one and install the same version of operation system(solaris 7).

I have backed up the first system with:

<pre>

mt -f /dev/rmt/0cn rewind

ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t0d0s0
ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t0d0s3
ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t0d0s4
ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t0d0s5
ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t1d0s0
ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t1d0s1
ufsdump 0uf /dev/rmt/0cn /dev/dsk/c0t1d0s3
</pre>

and put everything in to same tape.
Below you can see :

<pre>

isgsi01(root)40: df –k

Filesystem kbytes used avail capacity Mounted on
/proc 0 0 0 0% /proc
/dev/dsk/c0t0d0s0 1987399 936298 991480 49% /
fd 0 0 0 0% /dev/fd
/dev/dsk/c0t0d0s3 962571 645814 259003 72% /var
/dev/dsk/c0t0d0s4 962571 766780 138037 85% /home
/dev/dsk/c0t0d0s5 3009327 2066073 883068 71% /u01
/dev/dsk/c0t1d0s0 3009327 2262543 686598 77% /u02
/dev/dsk/c0t1d0s1 3009327 2323693 625448 79% /u03
/dev/dsk/c0t1d0s3 3009327 2938899 10242 100% /u04
swap 2761080 6592 2754488 1% /tmp

</pre>

The problem is that when I try to restore the (root) / and /var
directories the system core dumps and gives me a bus error.
I went through SUN and "ufsrestore procedure for restoring / and /usr" but it says somewhere that I have to unmount the
file system and create a new file system with newfs.
I can NOT do the umount for / or /var.
I get that the / or /var are busy or already mounted.
What can I do for this ?? I am in / directory when I do that.

I use 'ufsrestore srv 7' for example for the /u04 partition
and it works fine. It does the restore without problems.
The same with all the rest apart from /var and /.
Can anyone help please ??

Thank you.

If you need any other information about the ufsdump or restore
please let me know. I need to provide a full restore of the system
by today. Please help if you can !

I hope the below info help a bit.

I ufsdump the filesystems one by one using the SAME TAPE.
when you ufsrestore you can always say:

ufsrestore si 4 and would mean the /u01 partition.

so the file systems in one tape are numbered

1 for /
2 for /var
3 for /home
4 for /u01

etc

<pre>
Filesystem kbytes used avail capacity Mounted on
dev/dsk/c0t0d0s0 1987399 937764 990014 49% /
/dev/dsk/c0t0d0s3 962571 654355 250462 73% /var
/dev/dsk/c0t0d0s4 962571 727526 177291 81% /home
/dev/dsk/c0t0d0s5 3009327 2066417 882724 71% /u01
/dev/dsk/c0t1d0s0 3009327 2263831 685310 77% /u02
/dev/dsk/c0t1d0s1 3009327 2323695 625446 79% /u03
/dev/dsk/c0t1d0s3 3009327 2939427 9714 100% /u04

/proc 0 0 0 0% /proc
fd 0 0 0 0% /dev/fd

swap 2806960 9336 2797624 1% /tmp

</pre>

Thanks very much in advance.

The reason you were getting those errors is probably because ufsrestore was trying to restore stuff to /proc or some other realtime OS specific directories which are constantly changing (like /dev/kmem and so on).

Not a good idea to overwrite system memory and process table entries while the system is up

You would have to either restore only what you need (use interactive ufsrestore) or just restore it to another filesystem other than root (e.g, another disk somewhere)

Hello doeboy,

What I am trying to do is that I want the two systems to be exactly the same. Is this possible ??
Another quiestion: How do I set up te automounter is Solaris 7??

Thank you

If you want the systems to be exactly the same, then boot the target machine from a CD or something like that... then mount the disk, and do the ufsrestore to it.

Checkout automount(1M) and automountd(1M) man pages for more info. You'll have to create an /etc/auto_master if one isn't already there.