[c-nsp] FW: ME3600 FCS packet generation issue

[Jordi Magrané Roig]

Hello Ytti,

The connection is:

7609 ---- GigEth (Fibre) ---- ME3600-01 (with the issue) ------- EoSDH Network ------- ME3600-02

The issue is when the ME3600-01 forwards a packet in the 7609 port I see "input errors". In the ME3600-01 I also have customer devices, I manage some of these customer devices. In these customer devices I also see "input errors". The show controllers ethernet-controller command in the ME3600-01 reveals FCS errors. Reloading the ME3600-01 fix temporarily the issue.
In the ME3600-02 I don't see errors because there is the EoSDH Network in the middle. The EoSDH is Alcatel Lucent. The issue appeared when we connected the two ME3600 via the EoSDH Network.  The packets from a customer device connected to ME3600-01 and forwarded to 7609 also had errors. 

I'm not sure if the issue is the EoSDH network, I will check the Lucent card revision. I think that due to some event the device goes "crazy" and send packets with errors.

If you need more clarifications feel free.

Best regards,
Jordi.  


> Date: Fri, 24 Oct 2014 13:25:17 +0300
> From: saku at ytti.fi
> To: cisco-nsp at puck.nether.net
> Subject: Re: [c-nsp] ME3600 FCS packet generation issue
> 
> On (2014-10-24 10:12 +0200), Jordi Magrané Roig wrote:
> 
> Hi Jordi,
> 
> > ME3600. I have a ME3600 connected to a C7609 via the G0/24 port and 
> > connected to another ME3600 connected via G0/22 port. In the middle of 
> > the connection between the two ME3600 there is a SDH network (Ethernet 
> > over SDH). I have a recurring issue in the ME3600, for some reason when 
> > the device receive a packet and forwards the packet, the device sends 
> > the packet with FCS errors, so the other end detects the "input error" 
> 
> Strange so, do I understand correctly
> 
> 7609 -> SDH -> ME3600 -> X
> 
> X experienced FCS error? So ME3600 has already received packet happily from
> SDH and sent forward?
> 
> We've seen firmware fault in Lucent GBE card in revision S1:7, rev 3 and rev 6
> work fine. This fault causes FCS occasionally to be incorrect, but you should
> then observe the FCS error in ME3600 itself.
> 
> It would seem to me much more likely that fault would be in SDH, since usually
> EoSDH does not transport FCS but optimizes it away. But again, if I understood
> your topology correctly, error should be visible in ME3600 counters in
> interface facing SDH, not in X.
> 
> -- 
>   ++ytti
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/