Customer has Main and DR sites. Main site has become POD1. DR will be migrated as POD2.
Each POD has a Internal FW which is the gateway for WAN traffic. External FW whic is default gateway for Intenret.
POD1 has two L3 outs. POD2 will also have two L3 out.
Till now invoking a DR event is a manual process:
- Main site is shut down
- DR FW interfaces are enabled and it takes over same IP address as of Main Site.
- ISP is involved and same IP is advertised from DR site.
- Standby compute and storage becomes operational in DR Site.
- So all the Main site IP are now on the DR site.
Assuming:
POD 1 is primary for NETWORK A and DR for NETWORK B
POD 2 is Primary for NETWORK B and DR for NETWORK A
Goals:
Enahance failover/DR process
Optimization of traffic for stretch VLAN.
Question:
1)
In MultiPOD, if we create additoanl node/interface profiles in the existing L3 out that corresponds to DR FW where DR FW IP are enabled with different IP address; use static route of higher admin distance. So that traffic from DR FW never flows out, but once DR is invoked, primary L3 out goes down and traffic from DR L3 out would automatically assumes.
Or Create additional L3 out in DR site for NEtwork A? Whats the best way?
This way we dont need to manully enable the FW interfaces. it improves DR process bit.
2) If a VM moves from POD1 to POD2, it is using L3 out from Primary site (tromboning). Whats the best way to to laverage local L3 out. Can GOLF play any role here? Keep in mind their L3 out is with FW. How will it impact the first point of standy L3 OUt?
*They dont want LISP.
Thanks
Nadeem
