3B2 Hard Drives

[Greg A. Woods]

In article <4346 at sactoh0.SAC.CA.US> eric at sactoh0.SAC.CA.US (Eric J. Nihill) writes:
>[ stuff about Maxtors vs. 3b2's ]

I tried sending mail to Eric about this, but it was bouncing.  Perhaps
he will see it if I post it here... :-)

[ On Tue Nov 20 12:37 (Re: "Maxtor Drives"), Eric wrote: ]
>   It seems that there is a hardware configuration problem
> between a 3B2 and a XT series Maxtor. This problem of being
> able to read the defect table one minute and not the next
> seems to be common with several XT series Maxtors on several
> 3B2's. 

Nope, it's a driver bug.

>   Are there any options on the drive that should be set to
> make it work with the 3B2? Jumper pins? Cut traces??

Yup, there's a patch (sorry I don't have the actual patch, since my
machine is down and out [lightning strike], but you should be able to
get the fix from AT&T).

I found this patch did not fully correct the problem, but only reduce
the hangs to a very short (1-5 secs.) time.  The drive(s) will always
re-cal. though, and yes if you get a bad-block when this happens, it
can start thrashing.

The only real fix would be to re-write the id interrupt handler, and
or speed up the machine.  (The patch is for 3.1v2.)

I've used a pair of Maxtor-1140's on my 3b2 after installing this
patch with no other problems at all.  This is with the drives
formatted out beyond their capacity (i.e. to 1024 cylinders).  I've
only had one or two bad blocks form on a drive during an operating
period of about 1 year, and the bad-block software took good care of
mapping it out.

I have had one second-hand 1140 go bad, and anther re-built model
require a second re-construction because the shippers appeared to have
used it as a football....
-- 
						Greg A. Woods

woods@{eci386,gate,robohack,ontmoh,tmsoft}.UUCP		ECI and UniForum Canada
+1-416-443-1734 [h]  +1-416-595-5425 [w]    VE3TCP	Toronto, Ontario CANADA