cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
253
Views
0
Helpful
1
Replies

Dual MSFC failover behavior

epeeler
Level 1
Level 1

Here's what I have:

6509 with dual sup 1 + dual msfc.

Msfc1 is secondary hsrp for vlan193

Msfc2 is primary hsrp for vlan193

If I take down the interface for vlan193 on msfc1, msfc2 is no longer able to communicate with anything on the vlan at layer 3. I can see entries for all the vlan193 nodes in the arp table on msfc2 but the router cannot ping any of those devices in the table, nor can anything on the subnet ping the router.

Sniffing the local port of a vlan193 node shows:

1. Inbound arp request from msfc2

2. outbound arp reply from node

3. Inbound icmp request from msfc2

4. outbound icmp reply from node.

So, when trying to ping from msfc2 I see the arp request and reply at the node. I also see the icmp request and reply at the node but apparently the packet either never reaches the router or is not recognized by it. Checking the routing table on msfc2 shows the network directly connected. Everyone's arp tables have the correct information in them. Msfc2 can ping it's own interface.

The instant I bring the interface back up on msfc1, msfc2 is again able to ping anything on the subnet. All routes look the same, arp tables the same, everything.

The only think I can think of is something with CEF between the two supervisors is causing a problem.

One other possibly pertinent piece of information:

When I initially configured routing for this vlan I did the msfc2 first. Everything was working fine at that point. Once I created the vlan interface on msfc1 is when this behavior started until I had the msfc1 interface fully configured and brought it up. Now, any time I take the msfc1 interface down, I lose Ip connectivity between the vlan and the msfc2.

Note also that from outside of the vlan193, you can ping the vlan193 interfaces on msfc2 even when msfc1 is down. So nodes on vlan2 (for example) can ping the 193.1 interface but nodes on vlan193 cannot. It only effects IP traffic on the local subnet.

I'm baffled. Any insight would be appreciated.

1 Reply 1

bryan.lofland
Level 1
Level 1

Try flushing the ARP table on the end node. It could be caching the MSFC1's info and therefore not knowing what happened. If this is an XP host then arp -a lists the cache and arp -d should delete it.

Thanks,

Bryan

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Review Cisco Networking products for a $25 gift card