Seems like some of the problems are with the Adaptec driver included with SUSE. Here is what we have found in our Singapore office...
Mainly, we had stick with a RedHat 8 installation on the opteron server. This is because SuSE 9 with the 2.4.21-193 kernel has a very new Adaptec driver whic has the following warning in the >> kernel's log message,
kernel: scsi0:A:0:0: DV failed to configure device. Please file a bug report against this driver.
This apparently is because the Adaptec driver on the 2.4.21-193 driver has this Domain Validation (DV) SCSI feature introduced in their newer driver. The Brownie box apparently has no such feature and is thus unconfigurable by the Adaptec HBA.
Also, we have tried creating filesystems on a RAID array with a single LUN
at 1.4 TB. We have tried both the Qlogic 2300 and LSI FiberChannel HBAs,
but when we create a filesystem on this LUN, (either ext3 or xfs), once we
cross the 1 TB boundary, the system starts receiving SCSI errors...
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2147483651
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2155872259
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2164260867
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2172649475
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2181038083
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2189426691
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2197815299
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2206203907
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2214592515
Mar 8 15:26:45 gfs000pc kernel: I/O error: dev 08:11, sector 2222981123
The HBA makes no difference, and things seem to work fine on RedHat 8....
Any suggestions ?
Kevin Gassiot
Advanced Systems Group
Visualization Systems Support
Veritas DGC
10300 Town Park Dr.
Houston, Texas 77072
832-351-8978
kevin_gassiot@veritasdgc.com
Kevin_Gassiot@veritasdgc.com
03/05/2004 02:21 PM
To
suse-amd64@suse.com
cc
Subject
[suse-amd64] Problems with SCSI Array with LUN larger than 1 TB
We are seeing the following problem with a scsi array attached via an
Adapted 29160 SCSI card. The array has a single LUN of 1.3 TB.
Is there a limit to the filesystem size in the 2.4.21-193-smp kernel of a
terabyte or less ?
Is this a problem with the card, or a kernel issue ?
Mar 5 15:29:16 ppc003 kernel: PCI: Enabling device 02:01.0 (0015 -> 0017)
Mar 5 15:29:16 ppc003 kernel: scsi0 : Adaptec AIC7XXX EISA/VLB/PCI SCSI
HBA DRIVER, Rev 6.2.36
Mar 5 15:29:16 ppc003 kernel: