[2972] in linux-scsi channel archive
Re: BusLogic problems...
daemon@ATHENA.MIT.EDU (Dan Hollis)
Thu Dec 18 16:28:41 1997
Date: Thu, 18 Dec 1997 13:05:09 -0800 (PST)
From: Dan Hollis <goemon@sasami.anime.net>
To: Steffen Grunewald <steffen@gfz-potsdam.de>
cc: root@olupot.astra.net, lnz@dandelion.com, rstagg@csc.com,
linux-scsi@vger.rutgers.edu
In-Reply-To: <199712181034.LAA17439@zrt2.gfz-potsdam.de>
On Thu, 18 Dec 1997, Steffen Grunewald wrote:
> "Dan Hollis writes"
> |> If you run them hard (e.g. raid0 striped array of several of them), they
> |> lock up hard, as in they require a power cycle to recover. Probably that
> |> Sparc2 can't run that seagate hard enough to crash it.
> |>
> |> But our news server can. (4 x ST43400N striped with raid0).
> Does that lockup depend on the transmit rate ?
No. It has been tested on many different machines of completely
different configurations (486, pentium, amd), hostadaptors (ncr,
aic7xxx, buslogic), scsicables (various lengths and types), enclosures,
kernel versions (1.2.x through 2.0.33), driver versions.
All have the exact same result. The ST43400N locks up hard if it's run
synchronous.
> I had been running several Elite3's from a Sparc5 and a Sparc10, too.
> Never had lockups.
We have been running many ST43400N's for several years and have lots of
experience with them. They should be blacklisted for synchronous
operation.
> Perhaps limiting the tx rate would help here ?
Nope. Already been tested thoroughly.
-Dan
