HA path monitoring in virtual wire

[gmparis]

I've seen a couple answers here about using Path Monitoring in Virtual Wire. They say that one must use an IP address within the Virtual Wire subnet as the source address. OK, I get that. What I don't get is how to configure such an address. I don't see a way to add an address to a vwire interface. I've tried creating a loopback with no good result. Also gave vlan a shot, but that didn't look promising either. Thanks for any help.

[pkonitz]

As I described in this post

https://live.paloaltonetworks.com/thread/7403

I've got this release installed, got subinterfaces configured and assigned them to different zones but traffic doesn't flow if the main intarface isn't assigned to zone as well. When it is, subinterfaces inherits this assignment form main interface and even though they're in different one I see in logs that traffic comes from the main zone. So I cant make a different policies per subinterfaces.

Thanks for sharing your experience

I think we should end this topic cause gmparis will be angry for all the notification he receives

regards
Przemek

[Sly_Cooper]

I had asked permission from to use his thread so I hope he wont mind . Anyways I have submitted the change request to get the new config tested. If get approvals then I will get the config tested over the weekend. I will let you know the outcome.

[Sly_Cooper]

  - FYI

I successfully tested the new configuration and managed to configure path monitoring using additional link (L3) on PAN.

- Connected additional interface from PAN to the internal switch
- Configured L3 interface as part of the internal switch
- Configured new zone and virtual router 
- Configured new L3 interface as part of new zone and virtual router. This is mainly to keep the ha monitor link separate. I called it ha-monitor
- Configured Juniper firewalls with separate manage-ip for the untrust interface
- Configure Path Monitoring -> Virtual Router and monitored destination as the untrust manage-ip of the firewall
- Each PAN was configured to poll different manage-ip  of the connected Juniper firewalls
- Removed "Preempt" from HA (Thank you very much for the suggestion)

Failover and failback worked fine as expected and was tested by shutting down the firewall untrust port connected to the external switch. The separate manage-ip on Juniper firewalls and removing preempt made it work. Now I have link monitoring and path monitoring configured for our environment.

[NGS_SOC]

Hi, I've an installation  similar to your scheme and I uses PA-3020  under  ASAs in A/A with 4 vwire. Active Active configuration and full state sync allow you to forget witch ASA is active passing traffic.

[pkonitz]

hi,

thx for update

nice to hear it worked.

I decided to go to L2 deployment cause in Wwire mode passive unit doesn't pass traffic at all so in my case (cisco ASA) hello packets  did not flow so failover on ASA was a bit problematic. However L2 with vlan retagging works as charm, what is more, when ASAs switchover it does't trigger PAN to failover so I see additional benefit in it

As NGS said active/active probably would solve all our issues but this is thing I want to avaid cause guys from PaloAlto suggested not going into A/A

regards

[NGS_SOC]

They are always a little bit conservative with A/A design but I have a couple of big installation with A7A without specific issues. Also PA-3020 with PANOS 5.04 rulez in traffic response and commit time.

Regards

[gmparis]

I'm wondering why you used the manage-ip for the path monitoring instead of the NSRP IP. I have similar configurations here, both NSRP and VRRP, and my concern is that the shared address -- which is what the traffic cares about -- can be unreachable, even though the manage/local IP is reachable. This comes about because we generally have two layer-2 devices involved. This may not be your configuration, but still, I wonder if you are accomplishing what you want with this configuration.

Of course, I'm chiming in on this only because I'm getting all the updates to this thread. 🙂

[Sly_Cooper]

- Yes you may not see the issue with A/A firewall since both the firewalls will be passing traffic at a time. With Active/Passive you have to make sure that all connected devices in the path are passing traffic. Personally I am reluctant in configuring A/A model and I have never done that. A/A has its configuration complications. A careful configuration is needed. Since I was working on established A/P configuration and changing to A/A would need complete re-engineering.

- Technically I can use the NSRP ip however the issue with NSRP ip is that it becomes reachable for the failed firewall once the second firewall becomes Active. I have similar scenario as posted by where the ASA or Juniper firewall is egress/default route. The failed firewall can reach the NSRP ip via internal switch -> second PAN -> Juniper NSRP. Though this problem can be avoided by disabling Preemptive option as suggested by . I still wanted to make sure the PAN polls the correct interface for failure detection. Once the interface is UP, manage-ip too is UP. The manage-ip is available irrespective of the state of the firewall. Since Juniper provides the feature of configuring manage ip for interface,  I decided to use it. PAN is already taking care of connected devices failure detection using link monitoring (Juniper firewall trust and internal switch). I wanted to something to monitor untrust side of Juniper firewall as that is the ultimate point for traffic egress/ingress.