Loosing packets when running HSRP?

Unanswered Question
Sep 2nd, 2008
User Badges:


I have a network with 2 sites. On each site I have a stack ofh 3750:s. Between the sites there is a Layer2-connection from a provider. They deliver a 1Gbps l2-connection with ethernet-interfaces, so it works simply as one long patch-cable.

In the 3750:s there are 20 vlans created, and the l2-connection ports are simple trunks with no vlan filtering. The same vlans exists on both sites.

About 10 vlans are routed by the 3750 on site A. Each vlan-interface is configured for hsrp on both sites, where the 3750 on site A has higher priority in all vlan:s. So, all routing occurs at site A. And here's why:

If I create a new unused vlan on both sites and connect one host per site to the vlan (switchport access vlan x) i get a completely good working connection between them. When I create a L3-interface for the vlan on the switch in site B I can ping the router interface IP. However, when I configure the vlan interface for hsrp (only on site B at this moment) I can from the host on site A ping the IP-address dedicated to the vlan interface (int vlanx, ip address x.x.x.x) BUT when I ping the standby IP (int vlanx, standby ip x.x.x.y) i loose about 80% of the packets.


Site A


Host with IP .10 connected to 3750 switchport configured as "switchport vlan 400, switchport mode access".

A trunk connected to a switchport, going to site B. "switchport mode trunk"

Site B:


Host .3 connected to switchport configured as "switchport mode access, switchport access vlan 400"

A trunk connected to a switchport, going to site B. "switchport mode trunk"

3750 configured as

interface vlan 400

ip address .254

standby ip .1

standby priority ....

Now, on Site A:

.10 can ping .254 ok

.10 can ping .3 ok

.10 when pinging .1 gets 80% packet loss.

.3 can ping .1 ok

.3 can ping .254 ok

.3 can ping .10 ok

Also, when doing the opposite; creating the vlan400-interface ONLY on site A 3750 and trying to ping both IP:s from site B i am not getting any packet-loss. So it has something to do with running HSRP on the site B switch.

Any input? The switches are not running the exact same OS version, but they will do soon to eliminate this source of error.

Thanks in advance!

regards Jimmy

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
jilahbg Tue, 09/09/2008 - 21:59
User Badges:

Yes of course. But in what sense does it apply to my case?



glen.grant Wed, 09/10/2008 - 03:49
User Badges:
  • Purple, 4500 points or more

Maybe you can explain your setup a little more , don't understand why you have hsrp if you have a single link between the sites ?

jilahbg Wed, 09/10/2008 - 04:01
User Badges:


I have HSRP since I want routing to be with redundancy. The connection between the site is a L2 trunk. There is also WAN-connections that the ISP will deliver with HSRP to both sites.

Look at this as if there is only one site with a TP-cable-trunk between the 3750:s. Only that the 3750:s are separated with a looooong cable. ;-)


This Discussion