External disk problem.

[Bryn M. Reeves]

On Mon, 2012-07-30 at 18:04 +0800, Ed Greshko wrote:
> On 07/30/2012 05:59 PM, Bryn M. Reeves wrote:
> > On Mon, 2012-07-30 at 17:47 +0800, Ed Greshko wrote:
> >> On 07/30/2012 04:44 PM, Bryn M. Reeves wrote:
> >>> What state is the device in following a resume?
> >>> (/sys/block/sd*/device/state).
> >> I am not so sure that is a "good" indication of anything.  I have 2 drives on my system /dev/sda and /dev/sdb.
> > It's a fine indication of the state the kernel thinks the device is in
> > (that's what it's there for). If that file indicates the device is
> > running but in fact you can't issue I/O to it you'd suspect a problem in
> > the driver. If it's blocked or offlined you can look into what caused
> > that to happen.
> >
> >> [egreshko at meimei block]$ cat /sys/block/sdd/device/state
> >> running
> > Are they working? What are you trying to prove? For an active, working
> > device this is the normal state.
> >
> >
> 
> There is no disk plugged into the SATA port.   As I said....  I only have 2 sda and sdb.   There is no sdD.
> So, what is "running"? 

In this case a suggestion that for whatever reason the kernel hasn't
properly dealt with the removal of the device that had been assigned to
sdd (it presumably existed at some point for the sysfs path to have been
created).

If you're not getting entries in the logs when the problem happens you
can either look at other ways to get the logs out (e.g. serial console)
or you can try suspending and resuming from the console to see if
anything is printed (I have no idea if you are or not - you haven't
actually described the problem you are seeing so I'm just assuming it's
the "same" as Erik described since you are replying to his thread).

Bryn.