Hello,
I am trying to set up a failover system for our 60 Ghz wifi links. Most of these links are straightforward, but one is more complex.
The 60 GHz link is breaking up frequently during the autumn and I would like the fallback to be rapid.
The method which uses a ‘check gateway’ option on the route definition works OK, but takes up to a few minutes to activate the fallback connection.
I tried an alternative method using netwatch and enabling/disabling the route using a script (e.g /ip route enable [find where comment=“backup”]
In a lab setting this works fine, but on the production router it takes a long time - 15-30 minutes.
I expect this is because the router is a BGP router with nearly 1 million routes in the table.
This router is a CCR1009 router running ROS 6.42 - we are planning to upgrade this within the next weeks, but I am concerned performance will still be an issue.
Is there a smarter way to do this?
Thanks!
You still owe me a wireguard ( possibly using L2TP as well (not encrypted) dual WAN via CHR mt in the cloud setup help session for OSPF/BDF there LARSA.
Thanks for your reply.
I have not worked with OSPF before. It is a shared router, also running BGP, so I am a bit reluctant to try that. However, if it is the only solution I will study this option.
I think the delay is because the route I want to disable needs to be found (the find where comment=“backup” part). I expect this takes so long because we have close to one million routes stored.
Is it not possible to disable that route in a faster way?
Yes, route needs to be found before it can be disabled, and it is done by the console script, which adds to computing time even more.
v7 has faster lookup done directly by routing process, but with v6 there is nothing much you can do except reducing amount of routes.
@benw: BFD works just as well with BGP for a quick reroute if you lose the primary link. Manually rerouting with a script is like reinventing the wheel.
We are operating this network with three groups, each with their own /24 subnet. So the router only defines static routes to those subnets.
Under the /routes section we only have BGP defined.
I understand that BFD works as a supplement to another routing protocol. So I would have to configure OSPF, correct? It always seemed to be a bit of an overkill, just to manage the failover, but maybe it is time to bite the bullet.
No need to use OSPF if you’re already familiar with BGP. Both protocols use BFD in pretty much the same way to achieve fast failover between routes. Either way, give it a try - it’s not hard to set up. Otherwise, it works as usual: one route for the primary link and one for the backup. Just set attributes like local-pref for the faster link.
Then enable BFD for the specific sessions. You can change the tx/rx hello packet intervals separately if needed.
