On Thu, 24 Jan 2008, Guennadi Liakhovetski wrote:

> On Mon, 21 Jan 2008, Alan Stern wrote:
> 
> > On Mon, 21 Jan 2008, Guennadi Liakhovetski wrote:
> > 
> > > 2. Before, if such a reset happened when the disk was mounted it 
> > > definitely meant SCSI errors, most probably, a disconnect. Now (2.6.22 
> > > Debian kernel) these resets seem to be handled transparently - no SCSI or 
> > > fs errors, just a delay in data transfer and then it just continues. Is 
> > > it 
> > > indeed a (relatively) new feature?
> > 
> > New since about 2.6.10.  :-)
> 
> Hm, last kernel I used on this hardware with this hd was about 2.6.16, and 
> with that one I got disconnects and SCSI / fs errors with aborted 
> transactions.

The outcome of a reset depends on exactly what went wrong to cause the 
reset.  Some faults can be recovered and others can't.  I don't recall 
there being any changes in this area since 2.6.16.

> > > How normal is it?
> > 
> > On a well-working system it isn't normal at all.  Resets like that are
> > a sign of something wrong.  It could be at the electrical level (cable
> > connections or power levels), firmware level, media level (read/write
> > errors), or possibly elsewhere.
> 
> I just checked the same disk with another system, running the same 
> software. The hardware is quite different though. No errors there with the 
> same load pattern, that reliably produced errors on the first PC. Looks 
> like a host's problem?

Could be... a power-supply issue perhaps.  Or maybe a cable-connection 
issue.

Alan Stern

-
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to