Re: [Scst-devel] Triggering InfiniBand port state change

[Curtis <cu...@ti...>]

On 28/09/11 05:13, Bart Van Assche wrote:
> On Sat, Sep 24, 2011 at 2:26 AM, Curtis<cu...@ti...>  wrote:
>> Now, I can cause Windows to look again by replugging the IB cable, but
>> that's hardly a practical solution.  I've been toying with using
>> "ibportstate", but haven't got it to work yet.
>
> Have you already tried "ibportstate<hca>  <port>  reset" ? Although it
> reports an error message here, it seems to be sufficient to cause
> opensmd to take the port to the Down state and a few seconds later
> back to the Active state.

I have, as I mentioned in my original post... but the error messages and 
lack of results gave me the impression it wasn't helping -- my suspicion 
was because I didn't have a switch.  However, on another setup which 
does have a switch, we did get positive results.

This would, however, require configuring or determining the correct GID 
and port number so the startup script knew where to send the reset.

> # ibportstate 1 1 reset
> Initial PortInfo:
> # Port info: Lid 1 port 1
> LinkState:.......................Active
> PhysLinkState:...................LinkUp
> LinkWidthSupported:..............1X or 4X
> LinkWidthEnabled:................1X or 4X
> LinkWidthActive:.................4X
> LinkSpeedSupported:..............2.5 Gbps or 5.0 Gbps or 10.0 Gbps
> LinkSpeedEnabled:................2.5 Gbps or 5.0 Gbps or 10.0 Gbps
> LinkSpeedActive:.................10.0 Gbps

> ibwarn: [3657] mad_rpc: _do_madrpc failed; dport (Lid 1)
> ibportstate: iberror: failed: smp set portinfo failed

Yeah, I saw these errors, as well as complaints it was not a switch.

Right now, the problem seems to have gone away using trunk, but I am 
keeping this as a backup plan.

--
Curtis