Messages in this thread | | | Date | Tue, 7 Sep 1999 12:10:12 -0400 (EDT) | From | Chuck Mead <> | Subject | Re: Adaptec 2920A & fdomain.o (again) (Bugzilla #4670) |
| |
On Tue, 7 Sep 1999, Rik Faith spewed into the bitstream:
> On Mon 6 Sep 1999 15:57:26 -0400, > Chuck Mead <chuck@moongroup.com> wrote: > > banner: > > AHA Adaptec2920 BIOS Version 3.0 1994-1995 > [...] > > scsi0: <fdomain> No BIOS; using scsi id 7 > > scsi0: <fdomain> TMC-36C70 (PCI bus) chip at 0x6100 irq 10 > > scsi0 : Future Domain 16-bit SCSI Driver Version 5.50 > > scsi : 1 host. > > > > We have 3 of these cards with two of them in operation right now so I've got > > plenty to choose from for testing! > > When I sent you private email about this a week ago, I suggested that the > kinds of problems you are seeing are typical of incorrect termination or a > bad device. Incorrect termination doesn't always make the SCSI bus fail > completely. Instead, incorrect termination can make the SCSI bus fail for > certain regions of the disk and/or under certain kinds of load. I think > this is what you are seeing, especially with the SCSI timeouts you > reported.
Nope... the terminations are correct... these are previously running systems... not new installs... they have had numerous Linux installs... the problem arose with RH 5.2 and is still extant with 6.0. Also, I think it's important to note that if you do a web search for incidences of failure of this card you will find thousands of people complaining about the very same error and notably, the majority do not appear to have been resolved.
> I'm really sorry that I can't help more. On occasion, there are people who > report bugs like the ones you have reported. Most of the time, when those > people solve their problem, the problem is bad termination, a bad cable, or > a failing disk drive. The rest of the time, they get another controller, > and we never reach resolution of the problem.
I have a test machine in place which is going to be used to resolve this issue... I am deliberately *not* going to change the hardware on it so that we can fix this bug. The test machine has been running NT4.0 for the last two years 24x7 without a hitch (so plz don't ask me to start checking termination issues on it... NT would not run at all if there were termination problems).
> Since I can't duplicate failures like this on my test system, I suggest you > try another type of SCSI controller.
On my other "production system" (my wife's WS really) we've moved to an 2920C which uses the aic7xxx driver.
> The Future Domain chipset is very old, and most modern SCSI cards, even > cheap ones, will have better performance than the one you have now.
I agree it's old but a lot of people are still using it (including me and several others on this list). FWIW we run a small office and don't need huge SMP boxen to do our work so we really see no need to go out and spend a bazillion $'s on new hardware unless it's absolutely necessary.
Thanx for the thoughts!
-- Chuck Mead, CTO, MoonGroup Consulting, Inc. <http://moongroup.com> Mail problems? Send "s-u-b-s-c-r-i-b-e mailhelp" (no quotes and no hyphens) in the body of a message to mailhelp-request@moongroup.com.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |