Alan Cox writes:
>> May 19 13:05:59 production kernel: i2o/iop0: No handler for event
>> (0x00000020)
>> May 19 13:05:59 production kernel: i2o/iop0: No handler for event
>> (0x00000020)
>
> It sent us two event messages for events we've never head of (or asked
> for)
If we never asked for these messages, why did it send them?
>
>> May 19 13:06:04 production kernel: i2o/iop0: Hardware Failure: Unknown Error
>
> and then exploded
>
Yep. That's what's happening. Now, how do I defuse it? I *think* I'm setting
this up just as suggested on this list and various other sources, yet still
it is a no-go.
Btw. I got those error messages while trying to copy data from an
NFS-mounted volume from an old, slow, busy server to my RAID. This tells me
that probably the errors aren't caused by the SX6000 getting floodded by
incoming data stream and not being able to keep up with it.
On Maw, 2003-05-20 at 15:11, kernel wrote:
> > It sent us two event messages for events we've never head of (or asked
> > for)
> If we never asked for these messages, why did it send them?
Ask the vendor
> >
> >> May 19 13:06:04 production kernel: i2o/iop0: Hardware Failure: Unknown Error
> >
> > and then exploded
> >
> Yep. That's what's happening. Now, how do I defuse it? I *think* I'm setting
> this up just as suggested on this list and various other sources, yet still
> it is a no-go.
Ask the vendor. I've traced one of these with an end user and we didnt
send a single invalid I2O message to it
> NFS-mounted volume from an old, slow, busy server to my RAID. This tells me
> that probably the errors aren't caused by the SX6000 getting floodded by
> incoming data stream and not being able to keep up with it.
That would seem to be a bug in the controller firmware if so. You might
be able to reduce it or mitigate it by reducing the request queue size
we run.
Alan