OSPF Error message on E3500yl-48G

[altpes]

Hi,

One of ours E3500yl-48G ProCurve switches is constantly (every 10 sec. or so) generating same error message:

e 11/17/11 15:09:23 03132 OSPF: RECV: Discarding invalid packet : No such interface

Any idea what exatcly is wrong? And how to interpret that error?

Thank you

[ISoliman]

How is the OSPF configured? can you post the config and the diagram for this ?

[epes]

Hi,

Below is the config of the switch. It is a secondary switch that works with another one (ProCurve 8212zl). Just in case , at the bottom is the OSPF config of the primary switch. Appreciate your help. Thanks.

hostname "switch2"

ip routing

vlan 1
   name "DEFAULT_VLAN"
   untagged X,X,X
   ip address dhcp-bootp
   no untagged X,X,X
   exit

vlan 10
   name "Servers"
   untagged X,X,X
   ip address 10.1.1.3 255.255.255.0
   tagged x,X,X
   exit

ip route 0.0.0.0 0.0.0.0 10.1.1.1
ip router-id 10.1.1.3
router ospf
   area 0.0.0.1
   area backbone
   enable
   exit
router vrrp
   virtual-ip-ping
   enable
   exit

vlan 10
   ip ospf 10.1.1.3 area backbone
   vrrp vrid 10
      backup
      virtual-ip-address 10.1.1.2 255.255.255.0
      priority 192
      enable
      exit
   exit

###########################

hostname "switch1"

ip routing
vlan 1
   name "DEFAULT_VLAN"
   untagged X,X,X
   no untagged X,X,X
   no ip address
   exit
vlan 10
   name "Servers"
   untagged X,X,X
   ip address 10.1.1.2 255.255.255.0
   tagged Trk1-Trk5
   exit

ip route 0.0.0.0 0.0.0.0 10.1.1.1
ip route 10.1.5.0 255.255.255.0 10.1.1.11

router ospf
   area 0.0.0.1
   area backbone
   exit
router vrrp
router vrrp virtual-ip-ping

vlan 10
   ip ospf 10.1.1.2 area backbone
   monitor all both mirror 2
   monitor all both mirror 3
   vrrp vrid 10
      owner
      virtual-ip-address 10.1.1.2 255.255.255.0
      priority 255
      enable
      exit
   exit

[ISoliman]

why are you enabling VRRP and OSPF under the same vlan???  this vlan should be as a layer 3 interface, whst do you wsnt to accomplish exactly ? 

[epes]

VRRP is for redundancy and OSPF if for dynamic routing :) I guess. The person who configured it (and came up with the design) is gone (was a consultant) and we are not in the position at this point to re-design (have too many other projects running).

Do you think this can cause an issue?

[James_Levit]

What is area 1 for?

router ospf
   area 0.0.0.1

[epes]

Can't really explain - that is the way it was built. I simply inheritted that environment. I do not see any reason for separate area. Do you think that is what causing the issue?

[jguse]

Hello,

A few things I can point out about the configuration:

1) OSPF Config is odd, as others have mentioned. Whether area 0.0.0.1 should be there is hard to say without a proper topology including OSPF and VRRP design, and if you don't have that it wouldn't be surprising if the network was poorly designed and there are design-related issues. That error message seems to indicate a configuration/design issue too, and is definitely not any known software issue, especially considering how many people use OSPF on these switches. It would have been known by now.

2) ip router-id is only defined on Switch2, and missing on Switch1? What about a loopback interface, which should be defined as a best practice when using OSPF?

3) What are these entries being used for?

 monitor all both mirror 2
 monitor all both mirror 3

Is this for packet captures on the network, or is there some other reason for this configuration?

4) No VRRP Preempt Delay Time is set on the owner of VRID 10, Switch1. This is not good because of the following scenario. Switch1 goes down, VRRP fails over to Switch2 as owner right away, which is fine. Switch1 comes back up and because of its higher priority takes back VRRP ownership right away. This is problematic because OSPF will not have converged so quickly, and depending on the size of the network could take 15-30 seconds or more to converge. Thus all traffic would be dropped until that happens.

http://h40060.www4.hp.com/procurve/uk/en/pdfs/application-notes/How_to_configure_VRRP_Preempt_Delay_on_ProCurve_switches_Configuration_note_Dec_08_EMEA_Eng_A4.pdf

Either way I'd like to recommend hiring someone experienced to take a look at this. An HP Solution Architect, Partner or just some experienced Consultant could do it for you, that's up to you, if you do not feel comfortable making changes to the configuration yourself.

[ISoliman]

Nice Justin :)

Regards,
Islam

[jguse]

Hi Islam! Nice to see you around here too :)

[ISoliman]

LOL :D Nice you still remember me, Gr8 answer for the post btw :)

Kind Regards,

Islam

[epes]

Thank you, Justin

I totally agree that the design has to be re-visited and corrected. And I will try to address this (there are always higher priorities out there :)) I was just surprised not to find any interpretation for the error that I see in logs. It says "invalid interface".. which interface? Still a mystery for me...

But, once again, thank you for the tips and the link - great document. I do appreciate your comments.

Regards,

Evgueni