Subject: Re: daily crashes with 1.6.1
To: None <current-users@NetBSD.ORG>
From: Tim Middleton <x@Vex.Net>
List: current-users
Date: 07/06/2003 21:34:10
On July 6, 2003 01:40 pm, Patrick Welche wrote:
> mysteriously freeze was because the clever tagging of the new ahc driver
> played badly with the scsi drives *which came with write-back cache
> enabled*. All was well when this was switched off. (scsictl sd0 getcache)

I tried this today, and while some services on the box in question did freeze 
eventually, the box remained pingable and most services stayed up. Strange.. 
this at least is an improvement over the previous state where everything 
would freeze and the box was unpingable. 

Makes me wonder if there aren't multiple problems for us complicating each 
other. Perhaps ahc is messing us up, as well as NFS... all of the services 
that went down would have been accessing one particular NFS mounted 
partitions on a SCSI drive (the box being the NFS server, the partitions are 
physically available to it)... while services that stayed up do not touch NFS 
partitions; or at least not that particular partition. Coincidence? Perhaps 
time will tell.

-- 
Tim Middleton | Cain Gang Ltd | Christianity didn't [...] for 20 centuries
x@veX.net     | www.Vex.Net   | [...] shit Hallmark before a live studio 
audience.