regarding the above bug. If I were to execute the command " show ip bgp x.x.x.x/y longer-prefixes" do I run the risk of crushing my network as a result of this bug?
I am looking for a way to query the routing table for longer prefixes without causing an outage. We are working on upgrading the code to 5.1 but that will take time, we are talking about over 1000 routers. So in the mean time, am I safe running the BGP show command instead of the "show route"?
The problem can be triggered if the following CLI command is executed and the routing table of the corresponding address family is empty.
show route [ipv4 | ipv6 | vrf ] longer-prefixes <prefix>
When the issue is triggered the CPU usage of rib process of the requested address-family is increasing significantly and remains at a high load.
So, show route is safe as long as the output is not empty. If you hit this issue, you can recover by restarting the rib. Restarting the rib does not have impact to forwarding tw.
Restart the affected rib process of the corresponding address-family. E.g. process restart ipv6_rib