Problem
The tape drive on the system worked properly prior to an inplace upgrade to OpenServer 5.0.7. After the upgrade, I get the following errors from the system when trying to access it:
# tape status
tape: can't open '/dev/xct0': No such file or directory# tar tv8
tar: cannot open: /dev/rct0# cpio -vict -I /dev/rct0
cpio: cannot open </dev> for input.# dd if=/dev/rct0 of=foo
dd: cannot open /dev/rct0: No such file or directory
(error 2)However, I can properly access the tape drive using the SCSI device nodes for the tape (/dev/*Stp*). For example,
tape status /dev/xStp0tar tvf /dev/rStp0cpio -vict -I /dev/rStp0dd if=/dev/rStp0 of=foo Cause
During the upgrade, the default device nodes are reset to point to a cartridge tape device. This can be evidenced by viewing the appropriate device nodes.
Prior to the upgrade, the device nodes for both *ct0 and *Stp0 pointed to the same place:
crw-rw-rw- 5 root root 46, 8 May 2 2001 /dev/nrct0
crw-rw-rw- 5 root root 46, 0 May 2 2001 /dev/rct0
crw-rw-rw- 3 root root 46,128 May 2 2001 /dev/xct0crw-rw-rw- 5 root root 46, 8 May 2 2001 /dev/nrStp0
crw-rw-rw- 5 root root 46, 0 May 2 2001 /dev/rStp0
crw-rw-rw- 3 root root 46,128 May 2 2001 /dev/xStp0After the upgrade, the *ct0 devices were reset to use major number 10:
crw-rw-rw- 5 root root 46, 8 May 2 2001 /dev/nrStp0
crw-rw-rw- 5 root root 46, 0 May 2 2001 /dev/rStp0
crw-rw-rw- 3 root root 46,128 May 2 2001 /dev/xStp0crw-rw-rw- 4 root root 10, 8 Aug 13 2002 /dev/nrct0
crw-rw-rw- 4 root root 10, 0 Aug 13 2002 /dev/rct0
crw-rw-rw- 1 root root 10,128 Aug 13 2002 /dev/xct0Additionally, a portion of the tape configuration in the link kit may have been inadvertently removed. There may be missing/incorrect data the following Stp driver configuration files:
/etc/conf/node.d/Stp
/etc/conf/pack.d/Stp/space.h
/etc/conf/pack.d/Stp/device.dat Solution
This has been reported to SCO Engineering.
To work around the problem, you will need to remove your tape drive configuration and re-add it.
Run the following (as root)# mkdev tape- Select #1 "Configure a SCSI or Enhanced IDE tape drive" - Select #3 "Remove a SCSI Tape Drive" - Enter "0" to remove tape Unit 0.
NOTE: you may not see any tape drives configured, but enter "0" anyway.
- Type "y" to "Update SCSI configuration?" - Type "q" to "leave current string as is"
This will bring you back to the "SCSI or Enhanced IDE Tape Drive
Configuration Program" where you can quit out of the mkdev tape
routine.At this point, there may still be a spurious entry for this tape drive in your scsi configuration. To confirm, run
# mkdev .scsiand select "3. Display current SCSI configuration". If you do still see an "Stp" entry, make note of the "Host Adapter Type", "Adapter Number", "ID", "LUN", and "Bus". Then, select "2. Remove a SCSI device" and enter the information just gathered when prompted for it.
You should now be ready to re-run `mkdev tape` to add your tape
drive back into the system. If you are not sure of the correct
answers to provide when prompted for "Host Adapter Type",
"Adapter Number", "ID", "LUN", and "Bus", refer to the
/usr/adm/messages file and find your "tape" line from before
the system was upgraded. Once this is completed and the kernel is
relinked (as you exit mkdev tape), reboot the system to run with the
new kernel.Note: If you had multiple tape drives configured on the system, you may need to repeat the above procedure for additional drives.
If problems arise with linking your kernel due to any of the tape configuration files getting corrupted, please refer to the following technical article to manually fix your tape configuration: Technical Article 110973, "What files get out of sync after trying to unsuccessfully add a tape drive?"
See also
Manual pages: tape(C) tar(C) mkdev(ADM)TA110973, "What files get out of sync after trying to unsuccessfully add a tape drive?"