Originally posted by: IonutG.

Hi,
I have a cluster with 2 IBM pSeries 650 nodes with AIX 5.1.
I made a backup with smitty mksysb to tape on node 1 and I have tried to restore from this tape to node 2.
I followed this procedure when booting from tape on node 2:
1. Boot in service mode
2.Next I choose option "3 Start Maintenance Mode for System Recovery" then "4 Install from a System Backup"
3. I choose HDD for rootvg restoration.
4. Installation is working for 9% then I receive the following error:
"
Installing Base Operating System

If you used the system key to select SERVICE mode,
turn the system key to the NORMAL position any time before the
installation ends.

Please wait...

cannot open /dev/rmt0.1: No such device or address
Please mount volume 1 on /dev/rmt0.1
...and press Enter to continue
Approximate Elapsed time
% tasks complete (in minutes)
9 17 3% of mksysb data restored.
"

I have tried with different tapes but I have received the same error.
Any idea about this error?

Originally posted by: mmveiga

1) Did you check for hardware errors on both nodes ?
errpt -a

2) Did you try to read the contents of the tape in the node where you created ?
mt -f /dev/rmtX rewind
mt -f /dev/rmtX.1 fsf 3
restore -Tvf /dev/rmtX

3) Did you try to restore on a third machine ?

Regards,
Marcelo.

Originally posted by: IonutG.

Hi,
You are right - seems to be a problem with the tape but when I made the backup I didn't notice any error.

test /:mt -f/dev/rmt0 rewind
test /:mt -f /dev/rmt0.1 fsf 3
test /:restore -Tvf /dev/rmt0
Please mount volume 1 on /dev/rmt0.
Press the Enter key to continue.

restore: 0511-133 There is a data read error.: There is an input or output error.
Ignoring data and continuing.
restore: 0511-123 The volume on /dev/rmt0 is not in backup format.
Mount volume 1 on /dev/rmt0.
Press the Enter key to continue.

I will take another backup...
Thank you,
Ionut Gumeni

Originally posted by: mmveiga

Ionut,

Sorry my mistake, the correct restore command should be : restore -Tvqf /dev/rmt0.1
Have not done tape backups in years ...

Regards,
Marcelo.

Originally posted by: IonutG.

So, the right command sequence is
tctl -f /dev/rmt0.1 rewind
tctl -f /dev/rmt0.1 fsf 3
restore -Tqvf /dev/rmt0.1
witch return corect result:
New volume on /dev/rmt0.1:
Cluster size is 51200 bytes (100 blocks).
The volume number is 1.
The backup date is: Tue Mar 29 16:15:52 WET 2011
Files are backed up by name.
The user is root.
10735 ./bosinst.data
10647 ./image.data
0 ./usr/sys
.........................

This means that tape, tape drive and backup format are OK.
Somewhere else is the problem.
I have tried with a bakup made with RECOVER_DEVICE=yes and with RECOVER_DEVICES=no and in both situations I got the same error when booting from tape on node 2.

Best Regards,
Ionut Gumeni

Originally posted by: mmveiga

Ionut,

Do you have AIX running on node 2 to run the same commands to test the tape and check the error log ? It could be that tape drive is bad on that server.
Do you have a NIM master in your network ?

Regards,
Marcelo.