Looking for help with SLX-OS MCT cluster

[brhodgins]

Hi,

Humbly looking for help with an issue that I have spent many hours on. AI has been no help either. I've tried to make sense of the SLX config manual and I'm obviously missing something. I'm new to Extreme, but not to Avaya ERS or Cisco switching. I'm able to ping the peer IP from either switch, but have never got the peer state or the ICL tunnel to change from "down" to "up" .

I've tried everything that I could find to bring up a MCT cluster between two SLX-9140s.

SLX-9140-A 0/53 -> SLX-9140-B 0/53
SLX-9140-A 0/54 -> SLX-9140-B 0/54

Here is my config: 

DC1-A-SLX9140-B# sh clu
Cluster MCT-PAIR 200
================
Cluster State: Deployed
Client Isolation Mode: Loose
DF Hold Time: 3
Configured Member Vlan Range:
Active Member Vlan Range:
Cluster Control Vlan: 4090
Configured Member BD Range:
Active Member BD Range:
No. of Peers: 1
No. of Clients: 0

Peer Info:
==========
Peer IP: 10.255.255.1, State: Down
Peer Interface: Port-channel 200
ICL Tunnel Type: NSH, State: Down

Client Info:
============
Name Id ESI Interface Local/Remote State
---- -- --- --------- ------------------
DC1-A-SLX9140-B# show ip interface ve 4090
Ve 4090 is up protocol is up
Vlan is 4090
Hardware is Virtual Ethernet, address is fc0a.81f8.f001
Current address is fc0a.81f8.f001
Interface index (ifindex) is 1207963642
Primary Internet Address is 10.255.255.2/30 broadcast is 10.255.255.3
IP MTU is 9000
Proxy Arp is not Enabled
ICMP unreachables are always sent
ICMP mask replies are never sent
IP fast switching is enabled
Vrf : default-vrf
DC1-A-SLX9140-B# show running-config interface port-channel 200
interface Port-channel 200
mtu 9216
description MCT-PEER-LINK
no shutdown
!
DC1-A-SLX9140-B# show running-config interface ve 4090
interface Ve 4090
ip mtu 9000
ip address 10.255.255.2/30
no shutdown
!
DC1-A-SLX9140-B# show ip interface brief | include Ve
Ve 132 192.168.132.250 default-vrf up down
Ve 4090 10.255.255.2 default-vrf up up

Please, what am I missing? RASlog shows me nothing useful.

Cheers,

Brad

[Balaji]

Hi Brad, 

Looks like the port-channel 200 is missing the layer2 configuration. 

Since you are using the vlan 4090 as the control vlan for the cluster, you need to add the vlan 4090 to the port-channel 200. 

switchport

switchport mode trunk

switchport trunk allowed vlan add 4090.

[brhodgins · Accepted solution]

Thanks Balaji for your reply. I was finally able to get the MCT cluster to form by adding an IP address to each SLX. As a result routing has become an issue. I think that it has to do with duplicate vlan gateways set on each switch. I'm unfamiliar as to how Extreme handles vlan routing across the MCT pair. If I attempt to ping a vlan 132 host from SLX-A that is in it's arp table it works. However If I ping a vlan 132 host from SLX-A that has it's ARP entry on SLX-B it fails. If I down VLAN 132 and it's IP of 192.168.132.250 on SLX-B everything works via SLX-A (also has 192.168.132.250 as ve interface for vlan 132). Do i have to do something with "anycast" or proxy arp to have one vlan IP shared between both leaf members?

[Balaji]

could you provide the MCT and VLAN configuration from both the SLX-A and B. Also, "show cluster" output from both the SLXs.