OSPF Hello and dead timers
Hello IT Pro, I know that hello and dead timers need to match on the same network segment, otherwise OSPF peers will never reach the full state.. Do those settings need to match across the entire area though? Ie.. On other network segments in the same area.
In production deployments do you typically see these timers reduced intentionally? the standard 10 hello and 40 sec dead interval seems like forever if you are waiting for a broken routing table to "right" itself.
-Adam
@Adam-Tyler said in OSPF Hello and dead timers:
Hello IT Pro, I know that hello and dead timers need to match on the same network segment, otherwise OSPF peers will never reach the full state.. Do those settings need to match across the entire area though? Ie.. On other network segments in the same area.
Technically, they do not have to match for the entire area but it is better if they do. If not you're likely to get some unusual OSPF routing update issues if routers go up and down.
In production deployments do you typically see these timers reduced intentionally? the standard 10 hello and 40 sec dead interval seems like forever if you are waiting for a broken routing table to "right" itself.
You do not tend to see them changed unless there is a reason to. If you change them the best practice is to keep the hello/dead timer ratio (i.e. Dead timer is 4X the hello timer).
It's for the purpose of convergence. The reason not to set the 40 seconds closer in the ratio is that the router could be a in transient state or it was just a fluke drop in power...etc. If they are closer, it means that OSPF would have to run through the SPF algorithm more often, this means more resources used on the router.
@Adam-Tyler said in OSPF Hello and dead timers:
Hello IT Pro, I know that hello and dead timers need to match on the same network segment, otherwise OSPF peers will never reach the full state.. Do those settings need to match across the entire area though? Ie.. On other network segments in the same area.
Technically, they do not have to match for the entire area but it is better if they do. If not you're likely to get some unusual OSPF routing update issues if routers go up and down.
In production deployments do you typically see these timers reduced intentionally? the standard 10 hello and 40 sec dead interval seems like forever if you are waiting for a broken routing table to "right" itself.
You do not tend to see them changed unless there is a reason to. If you change them the best practice is to keep the hello/dead timer ratio (i.e. Dead timer is 4X the hello timer).
It's for the purpose of convergence. The reason not to set the 40 seconds closer in the ratio is that the router could be a in transient state or it was just a fluke drop in power...etc. If they are closer, it means that OSPF would have to run through the SPF algorithm more often, this means more resources used on the router.
