[8159] in linux-scsi channel archive
Re: recovery behaviour with 1 bad + 1 good drive (aic7xxx)
daemon@ATHENA.MIT.EDU (Ricky Beam)
Tue Feb 22 18:32:09 2000
Date: Tue, 22 Feb 2000 16:55:25 -0500 (EST)
From: Ricky Beam <jfbeam@bluetopia.net>
To: Dan Jones <djones@valinux.com>
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <38B2F87D.5AAB6C3B@valinux.com>
Message-ID: <Pine.LNX.4.04.10002221643480.12259-100000@beaker>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
On Tue, 22 Feb 2000, Dan Jones wrote:
>No comment on how responsive Mylex is, but I wouldn't depend upon
<grin>
I feel sorry for the millions of people using Mylex hardware.
>RAID to protect me on a system that gets SCSI parity errors. It
Well, I wouldn't buy a controller that cannot survive a parity error.
(Hello, retry the damned command.)
>isn't designed for that. SCSI parity errors are natures way of
>indicating that something needs to be corrected in your storage
>configuration. Of course, finding exactly what is wrong can be
... or IBM needs to better test their drives before selling them. I've
narrowed it down (I think?) to one out of six IBM drives. In LVD mode,
the controller locks up with this drive on the bus -- it's not active
mind you, just plugged into the shelf. If I force the bus to SE mode,
it can survive (sometimes) with that drive in the shelf. Take that
drive off the shelf and it works either way.
Note, I'm still going to blame Mylex for stupid firmware. My only
problem is in selecting the next vendor in whom to trust. (I don't
care about the cost of the card. I'll gladly spend money to keep from
having to fly/drive to Baltimore to fix the machine -- I'm in North
Carolina :-))
>a trial i.e cable, connector, termination, loading, etc. Until
>SCSI3 is completely implemented, the basic assumption is that
>transfers on the SCSI bus must be error free.
That's a good asumption, but that _is_ why there's parity. I rank this
up there with your car's engine cutting off because the EEC detected
a sparkplug didn't fire.
--Ricky
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
