2000-12-29 15:47:39

by George R. Kasica

[permalink] [raw]
Subject: SCSI Problems since upgrade from 2.2.16

Hello:

I'm running an HP DAT 4mm Autochanger here and since going to 2.2.17
and 2.2.18 I'm seeing failures when it attempts to unload the tape and
load a new one while backing up using BRU PE...utilizing the mt or mtx
commands as follows:

mt -f $DEV rewoffl 2>&1 >/dev/null

OR

/usr/local/bin/mtx -f /dev/sg1 eject 2>&1 >/dev/null


If I use the MTX command set to "manually" change the tapes all is
well....any thoughts on the cause or a fix...I don't think the
hardware is broken due to the fact it runs fine "manually" by doing
the mtx -f /dev/sg1 next commnand to load the next tape

Pertinent info below:

Information about installed SCSI devices

Attached devices:
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: SEAGATE Model: ST32550N Rev: 0021
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 03 Lun: 00
Vendor: HP Model: C1553A Rev: NS01
Type: Sequential-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 03 Lun: 01
Vendor: HP Model: C1553A Rev: NS01
Type: Medium Changer ANSI SCSI revision: 02

Bus 0, device 9, function 0:
SCSI storage controller: Adaptec AIC-7850 (rev 1).
Medium devsel. Fast back-to-back capable. IRQ 10. Master
Capable. Latency=32. Min Gnt=4.Max Lat=4.
I/O at 0xd000 [0xd001].
Non-prefetchable 32 bit memory at 0xdf000000 [0xdf000000].

SCSI support

CONFIG_SCSI = 1
CONFIG_BLK_DEV_SD = 1
CONFIG_CHR_DEV_ST = 1
CONFIG_BLK_DEV_SR = 1
CONFIG_CHR_DEV_SG = 1
CONFIG_SCSI_MULTI_LUN = 1
CONFIG_SCSI_CONSTANTS = 1
CONFIG_SCSI_LOGGING = 1


SCSI low-level drivers

CONFIG_SCSI_AIC7XXX = 1
CONFIG_AIC7XXX_CMDS_PER_DEVICE = (8)
CONFIG_AIC7XXX_RESET_DELAY = (5)


> Len: 20480
>
> Dec 23 01:24:37 eagle kernel: [valid=0] Info fld=0x0, EOM Current
> st09:00: sense key None
> Dec 23 01:24:37 eagle kernel: Additional sense indicates
> End-of-partition/medium detected
> Dec 23 01:24:37 eagle kernel: st0: Async write error (write) 7fffffff.
> Dec 23 01:24:40 eagle kernel: st0: File length 641863680 bytes.
> Dec 23 01:24:40 eagle kernel: st0: Async write waits 30037, finished
> 1304.
> Dec 23 01:24:42 eagle kernel: st0: Error: 28000002, cmd: 10 0 0 0 1 0
> Len: 0
> Dec 23 01:24:42 eagle kernel: [valid=0] Info fld=0x0, FMK EOM Current
> st09:00: sense key None
> Dec 23 01:24:42 eagle kernel: Additional sense indicates
> End-of-partition/medium detected
> Dec 23 01:24:42 eagle kernel: st0: Buffer flushed, 1 EOF(s) written

===[George R. Kasica]=== +1 262 513 8503
President +1 206 374 6482 FAX
Netwrx Consulting Inc. Waukesha, WI USA
http://www.netwrx1.com
[email protected]
ICQ #12862186


2000-12-29 19:18:59

by Douglas Gilbert

[permalink] [raw]
Subject: Re: SCSI Problems since upgrade from 2.2.16

"George R. Kasica" wrote:
>
> Hello:
>
> I'm running an HP DAT 4mm Autochanger here and since going to 2.2.17
> and 2.2.18 I'm seeing failures when it attempts to unload the tape and
> load a new one while backing up using BRU PE...utilizing the mt or mtx
> commands as follows:
>
> mt -f $DEV rewoffl 2>&1 >/dev/null
>
> OR
>
> /usr/local/bin/mtx -f /dev/sg1 eject 2>&1 >/dev/null
>
> If I use the MTX command set to "manually" change the tapes all is
> well....any thoughts on the cause or a fix...I don't think the
> hardware is broken due to the fact it runs fine "manually" by doing
> the mtx -f /dev/sg1 next commnand to load the next tape
>
> Pertinent info below:
>
> Information about installed SCSI devices
>
> Attached devices:
> Host: scsi0 Channel: 00 Id: 01 Lun: 00
> Vendor: SEAGATE Model: ST32550N Rev: 0021
> Type: Direct-Access ANSI SCSI revision: 02
> Host: scsi0 Channel: 00 Id: 03 Lun: 00
> Vendor: HP Model: C1553A Rev: NS01
> Type: Sequential-Access ANSI SCSI revision: 02
> Host: scsi0 Channel: 00 Id: 03 Lun: 01
> Vendor: HP Model: C1553A Rev: NS01
> Type: Medium Changer ANSI SCSI revision: 02

While I am not familiar with mtx and the process you
are having problems with, 'man 1 mtx' contains the
following:
The first argument, given following -f, is the SCSI
generic device corresponding to your media changer.

On the basis of the /proc/scsi/scsi output you have shown,
the mtx commands should read "mtx -f /dev/sg2 ..."
(not /dev/sg1) as you have noted at the top.

Given those 2 sg devices are closely coupled (just
differing by the lun) mtx probably can sort this out.

Doug Gilbert