LLDP forewarding and profinet

[p.nordlien]

Dear all.

I have an unmanaged switch, and I am trying to replace it with this managed Cisco switch:

Model: IE-4000-8T4G-E
SW ver: 15.2(2)EA2
SW image: IE4000-UNIVERSALK9-M

When connect the three hosts that I have to the Cisco switch, communication stops working. It is somewhat critical because there is a PLC that controls water level in a tank through two pumps. Communication between the PLC and pumps are provided by this unmanaged switch and as soon as I connect them to the Cisco switch, the PLC loose communication with the pumps. I have done some Wireshark'ing on the unmanaged switch, and it seems like the Cisco switch blocks some critical LLDP Multicast, like this:

Frame 19: 60 bytes on wire (480 bits), 60 bytes captured (480 bits) on interface 0
Ethernet II, Src: Siemens_14:58:dd (00:1b:1b:14:58:dd), Dst: LLDP_Multicast (01:80:c2:00:00:0e)
    Destination: LLDP_Multicast (01:80:c2:00:00:0e)
    Source: Siemens_14:58:dd (00:1b:1b:14:58:dd)
    Type: PROFINET (0x8892)
PROFINET acyclic Real-Time, Delay, ID:0xff40, Len:  44
    FrameID: 0xff40 (0xFF40-0xFF43: Acyclic Real-Time: Delay)
PROFINET PTCP, DelayReq: Sequence=25436, Delay=0ns
    Header: Sequence=25436, Delay=0ns
    DelayParameter: PortMAC=00:1b:1b:14:58:dd
        TLVHeader: Type=DelayParameter (6), Length=6
        PortMACAddress: Siemens_14:58:dd (00:1b:1b:14:58:dd)
    End
        TLVHeader: Type=End (0), Length=0

As you can see, the communication between the hosts are profinet. I have tried to enable profinet with the global command on the vlan in question, but it makes no difference. I have tried to disable IGMP snooping so that the switch would forward these LLDP packets, but no avail. I have tried igmp querier and static multicast MAC addresses, but nothing works.

Could it be that the code in the switch are written to disregard forwarding of all LLDP protocol considering the fact that LLDP are used between directly connected hosts? Does anybody have experience with this?

Any input appreciated.

Regards

[olgr]

Hello.

We do see something similar when we mix up IE and Office mode Switches. Connection from a PLC (luckily less critical) to 3550 to 3750 to 3560 works (yes, we know these switches do not speak profinet at all), but the connectivity is anyway not timecritical.

But now, when we setup another connection from a IE3000 to 3750 to 3560 it does not work. If we place an unmanage Switch before (unmanaged->IE3000->3750->3560) than it works too.

Must be something on the IE which would like to handle a signal from the PLC wrong (we try to disable profinet-no success). By the way - we run many IE3000 in production and if the traffic is only between these Switches all is fine. (without any specific config settings).

Looking forward to read some hints here.

Regards,

Oliver

[p.nordlien]

This thread led me to the solution:
https://supportforums.cisco.com/discussion/11458276/siemens-profinet-and-cat2960s

"I found a document where si explained that "probably" siemens host are using packet with 802.1p header and vlan 0."

I fooled around with the config and ended up with this interface config:

interface FastEthernet1/12
   switchport access vlan 33
   switchport mode Access
   switchport voice vlan 33
end

It Works! Nice to have solved it after having troubled with it for a week.