[NTLUG:Discuss] SCSI problems - (Was: aic7xxx: infinite...)

[Steve Baker]

Buddy Brannan wrote:
 
> *Hmm* I'm also having a problem with a CD-R. (SCSI, of course)
> 
> Some months ago, I got a 2X write, 4X read external CD-R, a JVC which is
> the same drive as one of the Teac drives. I was actually able to burn an
> ISO image with it once, and I don't remember if it was before or after I
> upgraded to 2.2.14.

I'd just upgraded to 2.2.14 - but I bought the CD writer the day AFTER
I upgraded - so I had never tried it with an earlier kernel.

> In any case, here are the symptoms. BTW, I even tried
> compiling in SCSI disc, CDROM, and generic support into the kernel rather
> than as modules, and I get the same symptoms:
> 
> cdrecord -scanbus picks the CDROM up correctly, identifies it, etc.; when I
> loaded the SCSI stuff as modules, it detected correctly at boot.

Ditto.
 
> If I try to mount a CD, or write to a CD, I will get an error that there is
> no media in the drive, even when there is. This error takes an age to come
> up, on the order of several minutes and sometimes close to a half hour.

Wow!  Well, I can reliably mount the CD - but I have never waited more
than a couple of minutes to see if it would recover.  Since the keyboard,
mouse and network were all completely un-responsive - and programs running
before I typed the command seemed to have frozen, I gave up and rebooted
almost immediately.

I'll try leaving it for an hour and see if it magically completes.
 
> I'm not sure if the drive is bad, the cable is bad, or the SCSI card is
> bad. The card is an Adaptec 2906 (AIC7XXX) card.

Well, it seems to me that if "cdrecord -scanbus" works - then it must
have suceeded at least enough to ask the drive what type it is and to
read back the results.  Hence the thing is *basically* OK - although
we could imagine an intermittant problem.

Perhaps your drive isn't terminated properly?

If it's an internal drive, it probably has a jumper to enable termination.

External drives sometimes have two connectors on the back - in which case,
you'll need a 'termination plug' in the second connector...to stop the
electrons from falling out I suppose :-)

> Any ideas which is the most likely culprit? If I had to guess, I'd guess
> the drive, but I have no way to know for sure.

Well, if your symptoms are similar to mine, it could be a kernel bug
in 2.2.14.

One thing I tried was this:

  cdrecord -v driver=scsi2_cd dev={whatever} -toc

(the 'driver=scsi2_cd' option forces the command to use a very
generic SCSI drive reader - rather than the one it thinks it
should use on the basis of auto-detection).

I found that running that command without the 'driver=scsi2_cd' would
lock my machine up - but running the scsi2_cd driver made that command
work (it reads the 'table of contents' off the driver).

That tells me that the SCSI interface must be working - but there is
something wrong with the autodetected driver.

-- 
Steve Baker                  http://web2.airmail.net/sjbaker1
sjbaker1 at airmail.net (home)  http://www.woodsoup.org/~sbaker
sjbaker at hti.com      (work)