Friends in many customers use a failover solution based on NetWatch.
Works only with more skilled labor.
We tested the function “check gateway” routing rules and works with a constraint, the switch to the second link only occurs if the hardware connected to it fail.
Example: An RB450 connected to two routers with links to the internet.
With the function “check gateway” gateway enabled in “A” with metric 1, and gateway “B” with metric 2, both are directly connected to the RB450.
If the link “A” stop working more connected to the router RB450 not stop the “check gateway” does not change the traffic to the link “B”, as tested by the gateway “check gateway” not unresponsive, but the link after the equipment.
I would suggest an adjustment in the function “check gateway”:
I would like the function “check gateway” enabled us to inform the target to be tested only by gateway entered in the field “check gateway”, so when the link “A” stop he would switch to the link “B” and when the target return to be reached by the link “A” he would again operate.
That would leave the failover much more simple and functional.
What do you think?
How about dynamic routing?
Do not understand the question.
Had some example?
you use a failover based on netwatch. Why not base it on routes. Make all devices knoe about all the other, and put weight to routes. Use ospf, bgp or something to make it dynamic, and its should be redundant by it self. Without a netwatch etc.
The scenario is connection between headquarters and branch to make a Lan to Lan between them using ADSL.
We’re having to make 1 EOIP and 1 PPTP.
One for each link and using NetWatch works, is more complex.
If we do not like it because it comes in the array are ADSL links.
1 has fixed ip and use another EOIP the ip is dynamic.
To test the link NetWatch and put ourselves when we have 2 EOIP Check Gateway.
What we would like is that the target check was configurable gateway.
That would solve the problem.
If you have a suggestion I would be grateful.
I don’t really understand you, but if you have the control of routing in both side. a dynamic protocol would be preferred. If you don’t have control of the routing, eg, connection to your ISP (without Global routing), the check gatway arp, is nice. (to pic the second if first is down)