SCSI resets - How do i identify what hangs off bus 3

[seanharrodine]

Hi all,

I have received the following SCSI reset/lbolts on my K460 this morning (running 11.11)....

Jun 15 06:48:57 shpe02 vmunix: SCSI: First party detected bus hang -- lbolt: 255632414, bus: 3
Jun 15 06:48:57 shpe02 vmunix: lbp->state: 7020
Jun 15 06:48:57 shpe02 vmunix: lbp->offset: 80
Jun 15 06:48:57 shpe02 vmunix: lbp->uPhysScript: 680000
Jun 15 06:48:57 shpe02 vmunix: From most recent interrupt:
Jun 15 06:48:57 shpe02 vmunix: ISTAT: 09, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 0000000e
Jun 15 06:48:57 shpe02 vmunix: lsp: 0000000000000000
Jun 15 06:48:57 shpe02 vmunix: lbp->owner: 0000000049c08200
Jun 15 06:48:57 shpe02 vmunix: bp->b_dev: cd0330c0
Jun 15 06:48:57 shpe02 vmunix: scb->io_id: 3b4cba6
Jun 15 06:48:57 shpe02 vmunix: scb->cdb: 12 00 00 00 80 00
Jun 15 06:48:57 shpe02 vmunix: lbolt_at_timeout: 255632314, lbolt_at_start: 255631814
Jun 15 06:48:57 shpe02 vmunix: lsp->state: 5
Jun 15 06:48:57 shpe02 vmunix: Zalon Fatal Error -- lbolt: 255632414, io_status: b90008c0
Jun 15 06:48:57 shpe02 vmunix: scratch_lsp: 0000000049ba7400
Jun 15 06:48:57 shpe02 vmunix: bp->b_dev: cb030000
Jun 15 06:48:57 shpe02 vmunix: scb->io_id: 3b4cba5
Jun 15 06:48:57 shpe02 vmunix: scb->cdb: 12 00 00 00 80 00
Jun 15 06:48:57 shpe02 vmunix: lbolt_at_timeout: 255633114, lbolt_at_start: 255630114
Jun 15 06:48:57 shpe02 vmunix: lsp->state: 205
Jun 15 06:48:57 shpe02 vmunix: Pre-DSP script dump [0000000044266000]:
Jun 15 06:48:57 shpe02 vmunix: 0900002c 07f00d40 80880000 00000020
Jun 15 06:48:57 shpe02 vmunix: 09000040 008009c0 80880000 ffffffe0
Jun 15 06:48:57 shpe02 vmunix: Script dump [0000000044266020]:
Jun 15 06:48:57 shpe02 vmunix: 09000014 07f00e80 80880000 ffffffe0
Jun 15 06:48:57 shpe02 vmunix: c0000004 006803c8 006807f8 80080000
Jun 15 06:48:58 shpe02 vmunix: SCSI: Resetting SCSI -- lbolt: 255632514, bus: 3
Jun 15 06:48:58 shpe02 vmunix: SCSI: Reset detected -- lbolt: 255632514, bus: 3

dmesg shows that it seems to be bus 3 at fault.......

SCSI: First party detected bus hang -- lbolt: 255632414, bus: 3
lbp->state: 7020
lbp->offset: 80
lbp->uPhysScript: 680000
From most recent interrupt:
ISTAT: 09, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 0000000e
lsp: 0000000000000000
lbp->owner: 0000000049c08200
bp->b_dev: cd0330c0
scb->io_id: 3b4cba6
scb->cdb: 12 00 00 00 80 00
lbolt_at_timeout: 255632314, lbolt_at_start: 255631814
lsp->state: 5
Zalon Fatal Error -- lbolt: 255632414, io_status: b90008c0
scratch_lsp: 0000000049ba7400
bp->b_dev: cb030000
scb->io_id: 3b4cba5
scb->cdb: 12 00 00 00 80 00
lbolt_at_timeout: 255633114, lbolt_at_start: 255630114
lsp->state: 205
Pre-DSP script dump [0000000044266000]:
0900002c 07f00d40 80880000 00000020
09000040 008009c0 80880000 ffffffe0
Script dump [0000000044266020]:
09000014 07f00e80 80880000 ffffffe0
c0000004 006803c8 006807f8 80080000
SCSI: Resetting SCSI -- lbolt: 255632514, bus: 3
SCSI: Reset detected -- lbolt: 255632514, bus: 3

I have attached a full ioscan.

Can anyone confirm please how i identify what is actually hanging off bus 3 ?

Am i correct in thinking a....
"ioscan -fnC ext_bus"
and then its the 2nd column that shows the bus ??

So, in my case, it would be 8/12 ??

There are no other system problems so this seems an isolated error and to my knowledge everything is terminated correctly.

Many thanks in advance for any replies.

Sean

[Duncan Edmonstone · Accepted solution]

Sean,

The key element to look for in here is:

bp->b_dev: cd0330c0


So to decode this:

The first 2 bits (cd) represent the major number of the device file - so cd is hex for 205:

#lsdev | grep 205
205 -1 stape tape


So it's a tape drive...

The next bits are Bus/Target/LUN as in BBTL in hex, so 0330 is

c3t3d0

Which looking at your ioscan output is the tape at HW path 8/12.3.0 (/dev/rmt/8m)

HTH

Duncan

[seanharrodine]

Hi Duncan,

Many thanks for the clarification of identifying which actual element on 8/12 complained.

Rgds,
Sean