ERROR: Command ioctl to rewind tape failed

[Galen Benson]

I get this error when running the following Ignite command: /opt/ignite/bin/make_tape_recovery -x inc_entire=vg00

Tried running it to two seperate tapes and can run a tar command against them, but always get this error with Ignite. The errors happen on a system running hpux 11.11 although I can run it successfully on 3 other servers.

Any thoughts?

[David Bellamy]

Which tape device were you using. ignite must be run with a no rewind device.

[Galen Benson]

From what I understand the ignite command defaults to 0mn for it's path. From the command I was using, I don't think it's trying to do anything with a rewind device. Am I incorrect on the command I used?

[David Bellamy]

I usually have to specify 0mn when i run the make_tape_recovery command.

[Bill Hassell]

I would never rely on a default tape drive. If it is wrong, that is, the default is 0m as it is for tar then this may be expected. The correct command would be:

make_tape_recovery -I -v -x inc_entire=vg00 -a /dev/rmt/0mn

Always specify exactly the drive required (-a). I always specify _I so that on restore, the UI pauses to ask about options. -v is always useful to see what is happening.

[Galen Benson]

Thanks Bill, but when I run the command you suggest, I get the same error message & the make_recovery completes unsuccessfully. Is there anything else I should be checking?

[Bill Hassell]

Yes, check syslog.log for any I/O errors. Also, what model tape drive is this? The error message is from a low-level system call and may be related to drive firmware and tape driver revision.

[Galen Benson]

The model of the tape drive is C5683A & when I look at the syslog, it's a lot of Greek to me:

SCSI: isrEscape Controller at 0/1/1/1.
SCSI: Parity error -- lbolt: 15914221, dev: cd033040
lbp->state: 30008
lbp->offset: ffffffff
lbp->nominalOffset: 270
lbp->Cmdindex: 16
lbp->last_nexus_index: 14
lbp->nexus_index: 15
uCmdSent: 1600dd80 uNexus_offset: 7d654
last lbp->puStatus [0000000041ed5634]:
ff031300 00031300 ff031300 00031300
next lbp->puStatus [0000000041ed5644]:
00031300 ff031300 00031300 00031300
From most recent interrupt:
ISTAT: 0a, SIST0: 91, SIST1: 00, DSTAT: 80, DSPS: 00000000
lsp: 0x0000000050e73800
bp->b_dev: cd033040
scb->io_id: 3000109
scb->cdb: 4d 00 42 00 00 00 00 00 ff 00
lbolt_at_timeout: 0, lbolt_at_start: 0
lsp->state: 4005
Jump Table entry [0000000041ed0300]: ff011300 00079000
lsp->puScript [0000000041ed1000]:
090000c0 24e73e00 0058c000 0900003f
25678380 005a0000 78370000 00000000
DSAtbl->host_iocb_index: 16
DSAtbl->host_iocb_addr: 7dd80
stored scratcha: 0xff031300
scratch_lsp: 0x0000000050e73800
c8xx_iocb [ffffffff90010b00]:
1600dd80 ff001300 00079000 bf031f00
00000001 0007dd60 0000000a 0007dd68
Pre-DSP script dump [0000000041ed1000]: 090000c0
Script dump [0000000041ed1004]:
24e73e00 0058c000 0900003f 25678380
005a0000 78370000 00000000 80080000
NCR chip register dump for: 0x400200a
00: SCNTL3: bf SCNTL2: 80 SCNTL1: 10 SCNTL0: da
04: GPREG: 0e SDID: 03 SXFER: 1f SCID: 47
08: SBCL: ae SSID: 83 SOCL: 09 SFBR: 80
0c: SSTAT2: 09 SSTAT1: 06 SSTAT0: 01 DSTAT: 80
10: DSA: 90010b00
14: MBOX1: 00 MBOX0: 00 ISTAT1: 00 ISTAT: 08
1c: TEMP: 00079000
24: DCMDDBC: 090000a0
28: DNAD: 24e73e20
2c: DSP: 0007900c
30: DSPS: 24e73e00
34: SCRATCHA: ff031300
38: DCNTL: a1 DWT: 00 DIEN: 7f DMODE: 4c
3c: ADDER: 00eece0c
40: SIST1: 00 SIST0: 00 SIEN1: 97 SIEN0: 8f
44: GPCNTL: 2f MACNTL: 00 SWIDE: 00 SLPAR: 00
48: RESPID1: 00 RESPID0: 80 STIME1: 00 STIME0: fc
4c: STEST3: 80 STEST2: 00 STEST1: 0c STEST0: 76
50: RESV50: 00 RESV51: c0 SIDL1: 00 SIDL0: 80
54: CCNTL1: 01 CCNTL0: 01 SODL1: 00 SODL0: 00
58: RESV58: 00 RESV59: 00 SBDL1: 00 SBDL0: 00
5c: SCRATCHB: 00030003
60: SCRATCHC: c0ffffff
64: SCRATCHD: 00079000
68: SCRATCHE: 90010c00
6c: SCRATCHF: 00078300
70: SCRATCHG: bf031f16
74: SCRATCHH: 0007d654
78: SCRATCHI: 0c01bf1f
7c: SCRATCHJ: 1600dd80
bc: SCNTL4: 00
PCI configuration register dump:
Command: 0157
Latency Timer: c0
Cache Line Size: 20
SCSI: Resetting SCSI -- lbolt: 15914321, bus: 3 path: 0/1/1/1
SCSI: Reset detected -- lbolt: 15914321, bus: 3 path: 0/1/1/1

Sorry for probably including too much data, but I'm not sure out of this what's relevant.

Any ideas?

[Sandman! · Accepted solution]

From the output of syslog it looks like there is something wrong due to the lbolt error (possibly the SCSI controller attached to the tape drive). Major number in "dev: cd033040" i.e. cd (hex) = 205 (dec) points to the SCSI tape

# lsdev 205

and minor number "033040" points to...

# find /dev -type c -exec ll -lart {} + | grep 033040

[Galen Benson]

Sorry it took me so long to get back to this & thanks for everyone's help. It appears that my notification from HP didn't show up.

[Galen Benson]

Problem solved