[ofa-general] How to recover from a bad MAD status (110) from lid 6

[Hal Rosenstock]

On Tue, Apr 7, 2009 at 4:01 PM, Chuck Baker <Charles.Baker at sun.com> wrote:
> Hi,
>
> I encountered an error while running load tests on RHEL5.2 OFED 1.4
> connected
> to SRP targets, and am wondering how to recover.
>
> The error's I'm seeing is an I/O failed with EIO I/O error messages, and my
> load
> generator failed.
>
> Since the failure, the srp_daemon reports
>
> srp_daemon -a -o -c -n -i mthca0
> 07/03/09 10:54:13 : bad MAD status (110) from lid 6

110 is ETIMEDOUT
 id_ext=0003ba0001005504,ioc_guid=0003ba0001005504,dgid=fe800000000000000003ba0001005506,pkey=ffff,service_id=0003ba0001005504,initiator_ext=0655000100ba0300
> id_ext=0003ba000100575c,ioc_guid=0003ba000100575c,dgid=fe800000000000000003ba000100575e,pkey=ffff,service_id=0003ba000100575c,initiator_ext=5e57000100ba0300
>
> rebooting the target and then rebooting the initiator has made no
> difference.

Sounds like some sort of network issue.

> Any ideas on how to resolve this would be appreciated.

Can you smpquery between initiator and target ? If so, what about perfquery ?

Have you tried rebooting your SM ?

-- Hal

> thanks
> chuck