Half a year ago I got an old Exabyte EXB-8505XL (IBM-version), which worked
fine with the NCR 53c810-controller. Then I replaced that with an
Acard AEC-6712UW adapter (atp870u driver). I cannot get this to work reliably,
it even freezes the system at some point.
Setup:
ECS K7S6A, Duron 800 MHz, 512 MB RAM, RedHat 7.3 updated with 2.4.20-kernel,
Matrox 550 graphics.
I rechecked the usual SCSI-things (cabling & termination), and found no
problems; other devices also work fine.
I reshuffled the PCI-cards so the SCSI-adapter doesn't share an IRQ.
I updated the Exabyte's firmware to rev. 808 (file 8sc-0808). For
this operation I had to move the unit to a spare system with an 53c810
as on this system:
2002.12.25-15:16:31--EXPERT 7.4.122 (MSVC 8.00 - 05/15/00)
ASPI hardware: ACARD AEC-671X Driver: DOS SCSIMGR 3.0 (HBA 0)
(5) CHECK (0x00.00.00) No Sense
(5) CHECK No additional sense information
(5) Sequential Access Device (0x00.00.00) No Sense
(5) Sequential Access Device No additional sense information
Warning - No tests found for any of the scanned devices.
(5) Re-scanning
(5) CHECK (0x00.00.00) No Sense
(5) CHECK No additional sense information
(5) Sequential Access Device (0x00.00.00) No Sense
(5) Sequential Access Device No additional sense information
(5) No device found
I couldn't get a different configuration EEPROM file on it:
2002.12.25-14:24:27--EXPERT 7.4.122 (MSVC 8.00 - 05/15/00)
ASPI hardware: SYM53C8XX Driver: SYMBIOS LOGIC (HBA 0)
(5) UNITATT (0x06.28.00) Unit attention
(5) UNITATT Not ready to ready transition, medium may have changed
(5) INQUIRY: 1, 2, EXABYTE, IBM-8505, 808R, (8505)
(5) File : A:\8SE-0010.BIN
(5) Converting A:\8SE-0010.BIN to .TAP file, please wait...
(5) Conversion successful!
(5) WRITE BUFFER 0x5 0 344064 3624
(5) WRITE BUFFER (0x03.26.02.65) Medium error, EOM
(5) WRITE BUFFER Parameter value invalid
(5) Firmware load Failed...
This may be because it's an IBM-version; current config version is 8se-0288.
What I see in /var/log/messages when I boot up the system, the read a bunch
of files (about 120 MB each) from tape in a loop with `dd':
----------------------------------------------------
Dec 25 16:27:26 duvel kernel: Linux version 2.4.20 (root@duvel) (gcc version 2.96 20000731 (Red Hat Linux 7.3 2.96-113)) #4 Fri Dec 20 15:05:53 CET 2002
...
Dec 25 16:29:31 duvel kernel: SCSI subsystem driver Revision: 1.00
Dec 25 16:29:31 duvel kernel: aec671x_detect:
Dec 25 16:29:31 duvel kernel: PCI: Found IRQ 11 for device 00:0b.0
Dec 25 16:29:31 duvel kernel: ACARD AEC-671X PCI Ultra/W SCSI-3 Host Adapter: 0 IO:e000, IRQ:11.
Dec 25 16:29:31 duvel kernel: ID: 3 TOSHIBA CD-ROM XM-6401TA1001
Dec 25 16:29:31 duvel kernel: ID: 4 YAMAHA CRW4416S 1.0h
Dec 25 16:29:31 duvel kernel: ID: 5 EXABYTE IBM-8505 808R
Dec 25 16:29:31 duvel kernel: ID: 6 HP T4000s 1.08
Dec 25 16:29:31 duvel kernel: ID: 7 Host Adapter
Dec 25 16:29:31 duvel kernel: scsi0 : ACARD AEC-6710/6712/67160 PCI Ultra/W/LVD SCSI-3 Adapter Driver V2.6+ac
Dec 25 16:29:31 duvel kernel: Vendor: TOSHIBA Model: CD-ROM XM-6401TA Rev: 1001
Dec 25 16:29:31 duvel kernel: Type: CD-ROM ANSI SCSI revision: 02
Dec 25 16:29:31 duvel kernel: Vendor: YAMAHA Model: CRW4416S Rev: 1.0h
Dec 25 16:29:31 duvel kernel: Type: CD-ROM ANSI SCSI revision: 02
Dec 25 16:29:31 duvel kernel: Vendor: EXABYTE Model: IBM-8505 Rev: 808R
Dec 25 16:29:31 duvel kernel: Type: Sequential-Access ANSI SCSI revision: 02
Dec 25 16:29:31 duvel kernel: Vendor: HP Model: T4000s Rev: 1.08
Dec 25 16:29:31 duvel kernel: Type: Sequential-Access ANSI SCSI revision: 02
Dec 25 16:29:31 duvel kernel: st: Version 20020805, bufsize 32768, wrt 30720, max init. bufs 4, s/g segs 16
Dec 25 16:29:31 duvel kernel: Attached scsi tape st0 at scsi0, channel 0, id 5, lun 0
Dec 25 16:29:31 duvel kernel: Attached scsi tape st1 at scsi0, channel 0, id 6, lun 0
Dec 25 16:29:31 duvel kernel: st0: Block limits 1 - 245760 bytes.
Dec 25 16:30:56 duvel ntpd[708]: time reset 2.700165 s
Dec 25 16:30:56 duvel ntpd[708]: kernel time discipline status change 41
Dec 25 16:30:56 duvel ntpd[708]: synchronisation lost
Dec 25 16:47:11 duvel kernel: st0: Error with sense data: Current st09:00: sense key Aborted Command
Dec 25 16:47:11 duvel kernel: Additional sense indicates Overlapped commands attempted
Dec 25 17:22:31 duvel kernel: scsi : aborting command due to timeout : pid 31392, scsi0, channel 0, id 5, lun 0 Read (6) 00 00 80 00 00
Dec 25 17:22:31 duvel kernel: r3b=20
Dec 25 17:22:31 duvel kernel:
Dec 25 17:37:31 duvel kernel: que cdb= 8 0 0 80 0 0 last_lenu= 8000 scsi : aborting command due to timeout : pid 31392, scsi0, channel 0, id 5, lun 0 Read (6) 00 00 80 00 00
Dec 25 17:37:31 duvel kernel: r3b=20
Dec 25 17:37:31 duvel kernel:
Dec 25 17:37:31 duvel kernel: que cdb= 8 0 0 80 0 0 last_lenu= 8000 SCSI host 0 abort (pid 31392) timed out - resetting
Dec 25 17:37:31 duvel kernel: SCSI bus is being reset for host 0 channel 0.
Dec 25 17:37:31 duvel kernel: SCSI host 0 channel 0 reset (pid 31392) timed out - trying harder
Dec 25 17:37:31 duvel kernel: SCSI bus is being reset for host 0 channel 0.
Dec 25 18:13:14 duvel su(pam_unix)[1490]: session opened for user root by bram(uid=500)
Dec 25 18:13:19 duvel kernel: Attached scsi CD-ROM sr0 at scsi0, channel 0, id 3, lun 0
Dec 25 18:13:19 duvel kernel: Attached scsi CD-ROM sr1 at scsi0, channel 0, id 4, lun 0
Dec 25 18:13:49 duvel kernel: scsi : aborting command due to timeout : pid 31393, scsi0, channel 0, id 3, lun 0 Mode Sense 00 2a 00 80 00
Dec 25 18:13:49 duvel kernel: r3b=20
Dec 25 18:13:49 duvel kernel:
Dec 25 18:14:19 duvel kernel: que cdb= 8 0 0 80 0 0 last_lenu= 8000 scsi : aborting command due to timeout : pid 31393, scsi0, channel 0, id 3, lun 0 Mode Sense 00 2a 00 80 00
Dec 25 18:14:19 duvel kernel: r3b=20
Dec 25 18:14:19 duvel kernel:
Dec 25 18:14:19 duvel kernel: que cdb= 8 0 0 80 0 0 last_lenu= 8000 SCSI host 0 abort (pid 31393) timed out - resetting
Dec 25 18:14:19 duvel kernel: SCSI bus is being reset for host 0 channel 0.
Dec 25 18:14:19 duvel kernel: SCSI host 0 channel 0 reset (pid 31393) timed out - trying harder
Dec 25 18:14:19 duvel kernel: SCSI bus is being reset for host 0 channel 0.
----------------------------------------------------
I got the suggestion that the SCSI tape drivers in the latest 2.4.x versions was
quite bad, so I tried 2.2.3 as well:
----------------------------------------------------
Dec 26 14:02:47 duvel kernel: Linux version 2.2.23 (root@duvel) (gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)) #2 Thu Dec 26 14:00:25 CET 2002
....
Dec 26 14:03:32 duvel kernel: st: bufsize 32768, wrt 30720, max buffers 5, s/g segs 16.
Dec 26 14:03:32 duvel kernel: aec671x_detect:
Dec 26 14:03:32 duvel kernel: ACARD AEC-671X PCI Ultra/W SCSI-3 Host Adapter: 0 IO:e000, IRQ:11.
Dec 26 14:03:32 duvel kernel: ID: 3 TOSHIBA CD-ROM XM-6401TA1001
Dec 26 14:03:32 duvel kernel: ID: 4 YAMAHA CRW4416S 1.0h
Dec 26 14:03:32 duvel kernel: ID: 5 EXABYTE IBM-8505 808R
Dec 26 14:03:32 duvel kernel: ID: 6 HP T4000s 1.08
Dec 26 14:03:32 duvel kernel: ID: 7 Host Adapter
Dec 26 14:03:32 duvel kernel: scsi0 : ACARD AEC-6710/6712 PCI Ultra/W SCSI-3 Adapter Driver V2.3+ac
Dec 26 14:03:32 duvel kernel: scsi : 1 host.
Dec 26 14:03:32 duvel kernel: Vendor: TOSHIBA Model: CD-ROM XM-6401TA Rev: 1001
Dec 26 14:03:32 duvel kernel: Type: CD-ROM ANSI SCSI revision: 02
Dec 26 14:03:32 duvel kernel: Vendor: YAMAHA Model: CRW4416S Rev: 1.0h
Dec 26 14:03:32 duvel kernel: Type: CD-ROM ANSI SCSI revision: 02
Dec 26 14:03:32 duvel kernel: Vendor: EXABYTE Model: IBM-8505 Rev: 808R
Dec 26 14:03:32 duvel kernel: Type: Sequential-Access ANSI SCSI revision: 02
Dec 26 14:03:32 duvel kernel: Vendor: HP Model: T4000s Rev: 1.08
Dec 26 14:03:32 duvel kernel: Type: Sequential-Access ANSI SCSI revision: 02
Dec 26 14:03:32 duvel kernel: Detected scsi tape st0 at scsi0, channel 0, id 5, lun 0
Dec 26 14:03:32 duvel kernel: Detected scsi tape st1 at scsi0, channel 0, id 6, lun 0
Dec 26 14:06:17 duvel ntpd[611]: time reset 3.505686 s
Dec 26 14:06:17 duvel ntpd[611]: kernel time discipline status change 41
Dec 26 14:06:17 duvel ntpd[611]: synchronisation lost
Dec 26 14:25:58 duvel kernel: st0: Error with sense data: [valid=0] Info fld=0x0, Current st09:00: sense key Aborted Command
Dec 26 14:25:58 duvel kernel: Additional sense indicates Overlapped commands attempted
Dec 26 14:50:33 duvel kernel: st0: Error with sense data: [valid=0] Info fld=0x0, Current st09:00: sense key Aborted Command
Dec 26 14:50:33 duvel kernel: Additional sense indicates Overlapped commands attempted
-------------------------------------------------------
So what's the solution? Similar operations on the T4000s work fine, which
suggests the Exabyte is at fault. But why did it work OK with the old
adapter then? In any case I assume that when a nasty SCSI-device can hang
up the whole system, the low-level driver must be failing, OR the adapter
hardware is faulty. Is it?
Bram Bouwens
|
