[Users] Error : Missing master SM in the discover fabric

[Chris Hunter]

If you have login for mellanox support website, they have a KB article
about this ie)
"Flashing Orange LED means that one or possibly more ports have just
received a symbol error.

Possible causes are:
• Bad cable
• Bad connection
• Bad connector. 
"
chris hunter

>On Wed, 3 May 2017 14:17:15, Hal Rosenstock wrote:
>
> Who is switch vendor ? Which LED on switch is on ? Is it the port LED ? Is
> it solid or flashing (orange) ?
> 
> Rebooting the Linux box causes the link to the switch to be renegotiated.
> That would complete before SM comes up on that Linux box so that port/link
> won't be active until that occurs.
> 
> You should not have to manually start opensm but there is a window until
> opensm is started and gets far enough to discover and activate the subnet
> which would turn the port LED from orange to green.
> 
> When ibdiagnet on that machine detects no master SM, is opensm already up
> and running ? That sounds consistent with solid orange LED on switch port
> LED and opensm not yet started.
> 
> When the subnet is in that state, is opensm running or has it not yet
> started ?
> 
> -- Hal
> 
> 
> On Tue, May 2, 2017 at 4:28 PM, Kenja, Krishna (kenjakt) <
> kenjakt at mail.uc.edu> wrote:
> 
> > I have an infiniband network up and running between two linux boxes
> > through a switch. But when I reboot the system on which the subnet manager
> > is running, the lights on the switch turn orange. Running 'ibdiagnet' on
> > the box with subnet manger showed me the following error.
> >
> >
> > -I---------------------------------------------------
> > -I- Bad Fabric SM Info
> > -I---------------------------------------------------
> > -E- Missing master SM in the discover fabric
> >
> > Running opensm or restarting opensm service fixes this, but I am unable to
> > understand why I am having this problem in the first place and restarting
> > the opensm service after every reboot is obviously not a practical
> > solution. Can some one please point me in the right direction as to what
> > the issue might be here.
> >
> >
> > Thanks
> > Krishna
> >
> > _______________________________________________