Problems reaching host from specific VLAN

[bern46]

Hello everyone,

We currenlty have a stack of two Cisco Meraki, with a total of 3 VLANS configured. 

Port 3 of the active Meraki is serving default Vlan1 and Vlan9

Port 4 of the Meraki is serving Vlan10

This 2 ports are going to a Core switch, first one as trunk and second one as access.

The following is the strange behaviour when I try to reach a host from Vlan 9 (from the meraki ping tools)

- Ping from Vlan9 to host of Vlan9: fine

- Ping from Vlan10 to host of Vlan9: fine

- Ping from Vlan1 to host of Vlan9: drop

Does someone know what could be happening ? 

Thanks.

[balaji.bandi]

how is your trunk config  on the Core switch, cisco default VLAN 1, check spanning treee config - make sure Core switch as spanning trree for root for vlan 1

or try creating new vlan other than vlan 1 and test it.

[bern46]

Note that Cores are also cisco.

The trunk port without encapsulation and the access port connecting from the Core to the meraki without STP.

I checked the spanning tree and the root is poiting to another port where there's another switch.

[Julio E. Moisa]

Hi

Are you doing loadbalance? you can create etherchannel between the switches and transport all the vlans through these 2 ports via Trunk interface.

Check  what switch is the root bridge for vlan 1.

Regards.

[Flavio Miranda]

Hi

 I failed picturing your scenario. If you have two ports connected to the same switches, one port should be blocked as per Spanning tree policy. If not, you are at risk of crash you network with a broadcast storm.

   But considering that it is the way you describded and all is good, from where to where you are pinguing?  You mentioned from the meraki ping tools but the target host is on the Core?

  Sounds to be as some tagging problem. Something is tag or not tag while should or should not.

[bern46]

Hi guys,

Changed the priority to set Core switch as root for Vlan 1 and nothing.

Meraki port is configured with Vlan1 as default and Vlan9 as allowed.

Core switch port has Vlan1 as native and Vlan9 as allowed, with default encapsultation 802.1Q

I have another location with the same structure and everything works fine.

Thanks

[Flavio Miranda]

But the STP would not be a problem for the communication.  The STP could prevent a problem in case of a loop.

 I beleive you might be facing some compatibilty issue.

 Merak may not undertand the Core or vice versa.

[bern46]

Hi Flavio,

It doesn't look like a compatibility issue. Both Core switches are also from Cisco and as I mentioned I have the exact same config in another place and it works fine.

Thanks anyways.

[Flavio Miranda]

Meraki is not exactly the same as Catalyst. 

 You can open a Cisco TAC for further investigate.  You can also share the configs here as off now we did not see anything.