Tagged aggregate link eats DHCP, ping works..?

[Jayadmin]

I've got two 1910 switches, A and B, connected via a 2-port fiber trunk.

Everything works fine except that the trunk seems to be eating DHCP packets.

On A's side, there is a DHCP server and all the workstations get their addresses correctly.

On B's side, none of the workstation get any reply from the DHCP server.

If I manually configure an IP address (or set up a rogue DHCP server) on one of the "B" workstations, I can ping the DHCP server and any of the  "A" workstations just fine -- which I guess means that ARP broadcasts go through correctly.

The trunk/aggregate link is a tagged member of VLAN 1 on both switches, and all the workstation ports are VLAN 1 untagged access ports.

I've disabled IGMP and DHCP snooping, and broadcast suppression is at ratio 100%. No DHCP relay is configured on either switch as everything is supposed to be in the same IP segment.

[Jayadmin · Accepted solution]

Ok... this seems to have mysteriously fixed itself.

No idea what did it, but now DHCP is working also, and both ends of the aggregated link (49-52) are still "VLAN 1-3 tagged PVID 1234 (dummy) link type trunk" while the workstation ports are "VLAN/PVID 1 untagged access" -- as they are supposed to be.