Description of problem:The new libata subsystem does not allow me to write to my
atapi tape drive.
Version-Release number of selected component (if applicable):Fc6 kernel
How reproducible: Constant
Steps to Reproduce:
mt -f /dev/nst0 erase # as root
causes this output:
Mar 24 21:15:29 bagend kernel: st0: Current: sense key: Illegal Request
Mar 24 21:15:29 bagend kernel: Additional sense: Invalid command operation code
Tape marks to be overwritten so tape can be overwritten from beginning. Also,
the backup script fails with the same output.
Saturday morning I booted into my old Fc5 partition and ran by backup script
that uses the ide-tape module. It was successful and I could read the data by
issuing "tar -jtf /dev/nht0".
I then booted into Fc6 running the latest kernel 2.6.20-1.2933. There the device
nst0 is now automatically created and linked to tape and tape_nst0. Using the mt
command I can rewind, retension, fsf (1-N), and check status with no problem. I
can also read the data on the tape with the above tar command.
I can not erase or write to the tape with the scuzzy driver.
When I forward space the tape it will move to the next file mark but when I
check the status it does not know where it is by block per this output:
# mt -f /dev/nst0 status
SCSI 2 tape drive:
File number=2, block number=0, partition=0.
Tape block size 512 bytes. Density code 0x45 (QIC-3095-MC (TR-4)).
Soft error count since last status=0
General status bits on (81010000):
EOF ONLINE IM_REP_EN
When I try to write to the tape I get this error:
# less /var/log/messages | grep -i st0
# Below are the error messages when I try to write to the tape
Mar 24 21:15:29 bagend kernel: st0: Current: sense key:
Mar 24 21:15:29 bagend kernel: Additional sense: Invalid
command operation code
Do I need to tell the st driver explicitly the tape block size? Do I use the mt
command to do that or is there another means? Or am I just the victim of scsi's
past inability to access tape drives on the same cable as another device as in a
cd / dvd ?
Is there a way to get more error output from the drive?
Can you post /var/log/dmesg from bootup of a working kernel and a non-working one?
Created attachment 151337 [details]
Output from dmesg for the kernel that write access works.
Created attachment 151338 [details]
Output from dmesg for the kernel that write access does not work
I think this bug is now the same as bug 150335.
I will no longer post information here unless it is requested. I have posted a
substantial amount of follow-up data under bug 150335.
I don't even have mt installed here. What package is it in?
*** This bug has been marked as a duplicate of 150335 ***