From owner-freebsd-scsi Tue Aug 26 12:24:43 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id MAA01636 for freebsd-scsi-outgoing; Tue, 26 Aug 1997 12:24:43 -0700 (PDT) Received: from fly.HiWAAY.net (dkelly@fly.HiWAAY.net [208.147.154.56]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id MAA01591; Tue, 26 Aug 1997 12:23:54 -0700 (PDT) Received: (from dkelly@localhost) by fly.HiWAAY.net (8.8.6/8.8.6) id OAA05878; Tue, 26 Aug 1997 14:23:45 -0500 (CDT) Date: Tue, 26 Aug 1997 14:23:45 -0500 (CDT) From: David Kelly Message-Id: <199708261923.OAA05878@fly.HiWAAY.net> To: dburr@POBoxes.com, freebsd-scsi@FreeBSD.ORG Subject: Re: AHA-2940AU (aic7860) and weird error messages on SCSI hd Cc: , , AIC7xxx@POBoxes.com, FreeBSD@POBoxes.com, Hardware@POBoxes.com, List@POBoxes.com Sender: owner-freebsd-scsi@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk Donald Burr writes: > > I just installed a SCSI disk into my machine. It's an older Conner 540MB > drive (CP3540), but this drive has been serving me well in another > machine. I have it partitioned into a ~150MB /home partition, a ~32MB > auxiliary swap partition, and two other partitions (one ~200MB and one > ~150MB) that are currently not in use. > > Anyway, after I installed and partitioned this disk, I started to get > these error messages both on my console (in bright white letters) and in > my syslogs. These messages do not cause a kernel panic or something, and > the system seems to run well despite them. Since seeing the first message > I've been making hourly backups of my /home partition and comparing them, > and all the data compares OK. Visually inspecting the files in /home > shows no errors either. > > What could be going on here? Any ideas? I have already checked SCSI IDs > and termination, and everything is A-OK. [snip] I have a 5x86/133-P75 MB which also caused troubles with an Adaptec 2940 (a very old 2940). The failures were similar, altho mine were often fatal. The solution was to either 1) disable the external MB cache or 2) use write-thru mode on the cache rather than write-back. If the system would stay running, write back cache was a couple of % faster than write-thru. Write-thru was about 10% faster than nothing. Guess how its been running the past year? :-) Another system didn't like tagged queuing. You might disable that and see how it works. My tagged queuing problem system looks like: ahc0 rev 1 int a irq 11 on pci0:15 ahc0: aic7860 Single Channel, SCSI Id=7, 3/8 SCBs ahc0 waiting for scsi devices to settle (ahc0:0:0): "IBM DPES-31080 !t S31K" type 0 fixed SCSI 2 sd0(ahc0:0:0): Direct-Access 1034MB (2118144 512 byte sectors) (ahc0:3:0): "MAXTOR 7213-SCSI 0742" type 0 fixed SCSI 1 sd1(ahc0:3:0): Direct-Access 202MB (415600 512 byte sectors) Had no problem until I partitioned the Maxtor and started using it. Dmesg didn't indicate the kernel was going to use tagged queuing on the Maxtor. Kernel simply paniced and rebooted. But that's another story. -- David Kelly N4HHE, dkelly@hiwaay.net (hm) ====================================================================== The human mind ordinarily operates at only ten percent of its capacity -- the rest is overhead for the operating system.