As Christopher T. Johnson wrote ...
> > As Christopher T. Johnson wrote ...
> > 
> > > REGARDLESS!
> > >   The drive is reporting an error condition and that error is NOT
> > >   getting back to user land.  I've verified the problem with amdump
> > 
> > Oh? Where/what is the error, does the console tell you? 
> 
> Sorry, no error reported, just the lights.  Amber light, no green solid.
> Either the kernel has the device "active" OR it is an error indicator.
> Exabyte is known for doing that.

In my experience it was always the drive, but YMMV.

> > >   amflush and dd.  In ALL cases once the amber light goes on solid
> > >   systat -vmstat shows a solid 1.4MB/second data transfer into
> > >   never never land.
> > 
> > That sounds like a bug. But it could be CAM not understanding that the
> > 8200 firmware went to Nowhere-land. If I got a dime for each Exabyte
> > lockup..
> 
> Yep, been there done that.  But these drives have been rock solid.  It is
> not the case of putting in a new, unknown drive and finding out that
> it is a little strange.  It is the case where I have a hardware configuration
> that has been stable that suddenly isn't.

CAM might trigger different code paths in the firmware. Or different timing.
Or.. <who knows>

> The original change is the upgrade to 4.0-CURRENT.  After that drive
> "failed" I took our back up drive and put it into production and got exactly
> the same failure mode.
> 
> Here is the sequence that I just did:
>       Old working tape.
>       mt erase &
>               1hr 30minutes or so later
>               Drive makes sounds of tape stopping.

I also tried that in the meantime (it said *click* just a minute ago ;-)
and I did not experience any problem with the erase operation.

>       Amber light comes on

Nope, after the tape motion stops I have the green one on again.

>       mt erase reports done (no errors)
>       mt rewind
>               mt: /dev/nrsa0: Device busy

p100#mt -f /dev/rst0 rewind
p100#

My P100 testbox uses a NCR810 to drive the 8200. 

Any chance you can try an 8200 with different f/w ?

I'll also try a 'dump' again with more data, my previous test was 300Mb.
Maybe that is not enough (I doubt it, but still..)

Will post results when it's finished writing.

Groeten / Cheers,

|   / o / /  _           Arnhem, The Netherlands        - Powered by FreeBSD -
|/|/ / / /( (_) Bulte    WWW  : http://www.tcja.nl      http://www.freebsd.org


To Unsubscribe: send mail to majord...@freebsd.org
with "unsubscribe freebsd-current" in the body of the message

Reply via email to