More Adaptec 29320 + Seagate ST336607LW woes

[Don Lewis]

On  5 Dec, ict technician wrote:
> On Thursday 04 December 2003 6:41 am, Justin T. Gibbs wrote:
>> Can you try to reproduce this issue with the driver posted here:
>>
>> http://people.freebsd.org/~gibbs/aic7xxx-20031203.tar.gz
>>
>> I found and fixed a few bugs that *may* be the root cause of your
>> issue.  I'm still trying to replicate this problem locally so that
>> I can independently verify the fix.
>>
>> Thanks,
>> Justin
> 
> Here are the results so far.
> 
> Firstly the box has run fault free all week.
> 
> I had squid hitting the default resource limits so I decided to put the increased
> values back in, before I started. Straight away I got a card dump. Typical!
> 
> I built a new kernel, removed the tags depth fix and enabled soft updates.
> It boots, it builds a new kernel, yahoo!
> 
> Now I thrash the living daylights out of the drive and it dies after 20mins or so, but
> frankly I'm not too bothered. I had to try really hard to kill it. This is just *so much*
> better.

I've been meaning to suggest that if you haven't done so already, use
"camcontrol modepage" to set the WCE bit to 0 which will disable write
caching on the drive.  In the testing that I've done, I haven't seen any
significant performance penalty in doing this if tagged command queuing
and softupdates are in use.  It is also safer for soft updates because
the drive no longer lies about when write data reaches stable storage.

Also, at least with Seagate drives, I have observed that the drive will
dynamically reduce the number of tagged commands that it will accept as
its write cache fills.  Who knows what sort of subtle bugs that might
tickle ...