[ofa-general] ib_mthca 0000:0d:00.0: Async event 16 for bogus QP 00da0407

[Bernd Schubert]

On Monday 06 April 2009, Tziporet Koren wrote:
> Bernd Schubert wrote:
> > Hello,
> >
> > I'm fighting (as usual) with some Lustre problems and I think this time
> > it is IB related. In the logs of some systems I see messages like these:
> >
> > ib_mthca 0000:0d:00.0: Async event 16 for bogus QP 00da0407
>
> This message means the driver get an asynchronous event from the HW for
> a QP that was already closed.

Sorry for my late reply, had been busy with too many issues.

Somewhere I have the IB specs Erez sometime ago gave me, I really need to find 
the time to read them (main issue, is that they have to much pages to print 
them and the pdf looks horrible on my ebook reader, so not really suitable for 
trains or airplanes...).

So at http://www.oreillynet.com/pub/a/network/2002/02/04/windows.html?page=2 I 
see a QP is a queue pair, which is used for communication between two systems. 

So this messages means the application closed the connection, but there was 
still something queued (e.g. on the other side) and the hardware received that 
after on this host the connection was already closed?

Btw, in the mean time I already figured out the reason for our IB problem - 
clients got out of memory and that somehow caused IB issues, I'm going to send 
another mail about that.

Thanks,
Bernd