[2904] in linux-scsi channel archive
Problem with DPT EISA Raid from 2.0.27 to 2.0.32
daemon@ATHENA.MIT.EDU (Anthony Rumble)
Sat Dec 6 18:37:20 1997
Date: Sun, 7 Dec 1997 10:28:37 +1100 (EST)
From: Anthony Rumble <smilie@ce.com.au>
To: linux-eata@i-connect.net, linux-scsi@vger.rutgers.edu
I have a problem that I can't seem to solve in a bug that occured
somewhere between 2.0.27 and 2.0.32
Since I have two types of SCSI adapters, I am not sure which
is at fault.. so I am sending it to both the authors/lists
Details of the system
ALR Revolution Q
Pentium 90
64 Megs of ram
EISA
2 SCSI adapters
Twin bus AHA-274x (EISA)
DPT Raid Controller (EISA)
The DPT has 3 hot swap drives in a raid array
The AHA has the Tape drive and 5 small 500 meg SCSI drives
split across the duel scsi busses, using MD to stripe them togther
as a fast scratch drive.
In kernel 2.0.27 we had no problems whatsoever.
Now that we have upgraded to 2.0.32 we have problems..
First.. we get these messages often
Dec 7 08:49:12 enterprise kernel: aic7xxx: Command complete near Qfull count, q
outcnt = 3.
I have no idea what this means..
Second.. and more importantly..
During normal operation, the system runs fine (other than the messages above)
BUT. when the nightly backups run.. most of the time the system falls over.
The tape backup unit is on the AIC controller, but it's only backing
up data on the DPT (The drives on the AIC are scratch drives so we don't
back them up)
It generally happens half way through the backup
I don't have a copy of the errors on the screen yet when it crashes, I am
still waiting for the Sysops to send me the transcript.. But they are
EATA errors, something like timeouts and card not responding and stuff..
Unfortunately.. it doesn't log anything once it happens..
Below are the boot up logs that may be of help.. When I get the
transcript of the error messages I will pass them onto you
ASAP.. I am sending you this now in the hope that you may allready
know what this problem is..
I have turned off command queuing, as I have had problems with that
on other machines/controllers.. but it didn't seem to help.
(elevator sorting is turned on tho)
Dec 6 20:56:09 enterprise kernel: aic7xxx: <Adaptec AHA-274X SCSI host adapter>
at EISA 5
Dec 6 20:56:09 enterprise kernel: aic7xxx: BIOS enabled, IO Port 0x5c00, IRQ 11
(level sensitive), Revision >= E
Dec 6 20:56:09 enterprise kernel: aic7xxx: Twin Channel, A SCSI ID 7, B SCSI ID
7, 4/4 SCBs, QFull 4, QMask 0x7
Dec 6 20:56:09 enterprise kernel: EATA0: address 0x5c88 in use, skipping probe.
Dec 6 20:56:09 enterprise kernel: EATA0: 2.0C, EISA 0x7c88, IRQ 15, BMST, SG 64
, MB 64, tc:n, lc:y, mq:8.
Dec 6 20:56:09 enterprise kernel: EATA0: SCSI channel 0 enabled, host target ID
7.
Dec 6 20:56:09 enterprise kernel: EATA/DMA 2.0x: Copyright (C) 1994-1997 Dario
Ballabio.
Dec 6 20:56:09 enterprise kernel: scsi0 : Adaptec AHA274x/284x/294x (EISA/VLB/P
CI-Fast SCSI) 4.1/3.2
Dec 6 20:56:09 enterprise kernel: scsi1 : EATA/DMA 2.0x rev. 3.11.00
Dec 6 20:56:09 enterprise kernel: scsi : 2 hosts.
Dec 6 20:56:09 enterprise kernel: scsi0: Scanning channel A for devices.
Dec 6 20:56:09 enterprise kernel: Vendor: CONNER Model: CFA540S
Rev: 0EAD
Dec 6 20:56:09 enterprise kernel: Type: Direct-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi disk sda at scsi0, channel 0, i
d 0, lun 0
Dec 6 20:56:09 enterprise kernel: Vendor: CONNER Model: CFA540S
Rev: 0EAD
Dec 6 20:56:09 enterprise kernel: Type: Direct-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi disk sdb at scsi0, channel 0, i
d 1, lun 0
Dec 6 20:56:09 enterprise kernel: Vendor: CONNER Model: CFA540S
Rev: 0EAD
Dec 6 20:56:09 enterprise kernel: Type: Direct-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi disk sdc at scsi0, channel 0, i
d 2, lun 0
Dec 6 20:56:09 enterprise kernel: Vendor: SONY Model: SDT-5200
Rev: 3.30
Dec 6 20:56:09 enterprise kernel: Type: Sequential-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi tape st0 at scsi0, channel 0, i
d 6, lun 0
Dec 6 20:56:09 enterprise kernel: scsi0: Scanning channel B for devices.
Dec 6 20:56:09 enterprise kernel: Vendor: CONNER Model: CFA540S
Rev: 0EAD
Dec 6 20:56:09 enterprise kernel: Type: Direct-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi disk sdd at scsi0, channel 1, i
d 0, lun 0
Dec 6 20:56:09 enterprise kernel: Vendor: CONNER Model: CFA540S
Rev: 0EAD
Dec 6 20:56:09 enterprise kernel: Type: Direct-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi disk sde at scsi0, channel 1, i
d 1, lun 0
Dec 6 20:56:09 enterprise kernel: Vendor: DPT Model: ENTERPRISE
Rev: 7C0
Dec 6 20:56:09 enterprise kernel: Type: Direct-Access
ANSI SCSI revision: 02
Dec 6 20:56:09 enterprise kernel: Detected scsi disk sdf at scsi1, channel 0, i
d 0, lun 0
Dec 6 20:56:09 enterprise kernel: EATA0: scsi1, channel 0, id 0, lun 0, cmds/lu
n 8, sorted, untagged.
Dec 6 20:56:09 enterprise kernel: scsi : detected 1 SCSI tape 6 SCSI disks tota
l.
Dec 6 20:56:09 enterprise kernel: SCSI device sda: hdwr sector= 512 bytes. Sect
ors= 1056708 [515 MB] [0.5 GB]
Dec 6 20:56:09 enterprise kernel: SCSI device sdb: hdwr sector= 512 bytes. Sect
ors= 1056708 [515 MB] [0.5 GB]
Dec 6 20:56:09 enterprise kernel: SCSI device sdc: hdwr sector= 512 bytes. Sect
ors= 1056708 [515 MB] [0.5 GB]
Dec 6 20:56:09 enterprise kernel: SCSI device sdd: hdwr sector= 512 bytes. Sect
ors= 1056708 [515 MB] [0.5 GB]
---
Anthony Rumble - Electronic Commerce Specialist
Corporate Express Australia Limited
Phone 02-9335-0669 Fax 02-9335-0753 Mobile 015-955-042 Pager 016-634-997
