I'm looking at implementing this command and I'm curious to know if this will affect OSPF neighbor adjacency's? I've looked through atleast 20 examples and I've never seen anywhere that shows by entering this command that it would drop OSPF neighbors...
the command influences how the interface ospf cost is calculated.
A change on this parameter should be done in all routers to assure consistency.
The sure effect of the command is that all ospf costs of the router are recalculated, if at least a change is experienced a new router LSA (actually one for every area the router is a member of) is generated and flooded in the network.
So for sure the deployment of the commands will cause a new generation of router LSA to be flooded within each area.
To be noticed that parameters used to build ospf adjacencies include:
area stub flags
ospf router id
but the cost of the link can be different for the two routers connecting to the same subnet and this doesn't block the ospf adjacency to form.
However, as correctly stated in another thread a test should be done.
I remember I've changed the auto-reference bandwidth several times in a lab environment and I've never noticed any change in the neighbor state and I was changing the parameter one router per one router.
A warning is shown that tells you to change the parameter in all routers in the domain or routing inconsistency can occur.
Unfortunately in these days I haven't access to a lab.
I would suggest to try it in a lab no downs are expected but practice rules.