Best practice for Active/Passive HA and OSPF

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Best practice for Active/Passive HA and OSPF

L1 Bithead

I configured Active/Passive HA in an environment where the firewalls connect to a core switch. There is an OSPF adjacency exists between the active Palo and the core switch. I'm curious what the best practice is for OSPF and HA. When tweaking the OSPF settings on the Palo, disabling OSPF graceful reset/strict LSA checking led to a vastly quicker failover. I cannot find any documentation on what the best practice is. Also, if anyone knows this, I would appreciate the theory on what exactly is supposed to happen during the HA failover with an OSPF adjacency. My understanding and hope is that the firewall that becomes active simply "continues" the OSPF adjacency, and that no new one needs to be formed. Or am I wrong here... Should the firewall actually form a new adjacency during a failover? 

1 accepted solution

Accepted Solutions

Cyber Elite
Cyber Elite

Hi @inssider ,

 

That is a great question.  The FIB is synchronized between A/P HA pairs, but the OSPF LSDB is not -> https://docs.paloaltonetworks.com/pan-os/10-1/pan-os-admin/high-availability/reference-ha-synchroniz....  You can confirm this by running the commands "show routing route" and "show routing protocol ospf lsdb" on the passive NGFW.

 

The purpose of OSPF GR is for the NGFW to tell its neighbor that there was a failover and to resynchronize the LSDB without taking the neighbor down.  OSPF GR is recommended for HA, but it should be enabled on inside router (OSPF neighbor) also -> https://knowledgebase.paloaltonetworks.com/KCSArticleDetail?id=kA10g000000Cm5ZCAS, at least in helper mode.

 

If the OSPF LSDB is empty, the NGFW probably will form a new adjacency to build it again.  I would check the system logs to see if OSPF flaps.  If failover takes longer with GR than without, I would check to see if it is enabled on both sides and check the timers.  The dead interval should not be less that the time needed for the passive control plane to take over.  Failover should be faster with OSPF GR than without, but use whatever works for you.

 

Thanks,

 

Tom

Help the community: Like helpful comments and mark solutions.

View solution in original post

2 REPLIES 2

Cyber Elite
Cyber Elite

Hello,

I have seen this in the past and just let the failover happen. OSPF has no issues with the failover even with default settings. Remember that the when the passive device becomes active, it takes the IP's of the former active so routing shouldnt be disrupted since the IP's are the same on the active/passive. While there might be small hiccups, I would just tune down the OSPF timers, but I'm running default and seldom have seen issues, etc.

 

Regards,

Cyber Elite
Cyber Elite

Hi @inssider ,

 

That is a great question.  The FIB is synchronized between A/P HA pairs, but the OSPF LSDB is not -> https://docs.paloaltonetworks.com/pan-os/10-1/pan-os-admin/high-availability/reference-ha-synchroniz....  You can confirm this by running the commands "show routing route" and "show routing protocol ospf lsdb" on the passive NGFW.

 

The purpose of OSPF GR is for the NGFW to tell its neighbor that there was a failover and to resynchronize the LSDB without taking the neighbor down.  OSPF GR is recommended for HA, but it should be enabled on inside router (OSPF neighbor) also -> https://knowledgebase.paloaltonetworks.com/KCSArticleDetail?id=kA10g000000Cm5ZCAS, at least in helper mode.

 

If the OSPF LSDB is empty, the NGFW probably will form a new adjacency to build it again.  I would check the system logs to see if OSPF flaps.  If failover takes longer with GR than without, I would check to see if it is enabled on both sides and check the timers.  The dead interval should not be less that the time needed for the passive control plane to take over.  Failover should be faster with OSPF GR than without, but use whatever works for you.

 

Thanks,

 

Tom

Help the community: Like helpful comments and mark solutions.
  • 1 accepted solution
  • 2961 Views
  • 2 replies
  • 0 Likes
  • 101 Subscriptions
Like what you see?

Show your appreciation!

Click Like if a post is helpful to you or if you just want to show your support.

Click Accept as Solution to acknowledge that the answer to your question has been provided.

The button appears next to the replies on topics you’ve started. The member who gave the solution and all future visitors to this topic will appreciate it!

These simple actions take just seconds of your time, but go a long way in showing appreciation for community members and the LIVEcommunity as a whole!

The LIVEcommunity thanks you for your participation!