HA-Groups - Monitoring separate pools on each LTM

 

Hi AllI have the following setup as depicted in the diagram attached:

 

I have two route domains running BGP b/w routers and F5s. The F5s are doing Route health injection.

 

I also have a HA-Group monitoring the trunk link b/w the F5s and the S1 and S2.

 

Recently, we realised that upstream work on R1 and R2 did not initiate sync failover (as the trunk link from F5 to switch stays up).

 

I have 2 traffic groups (one for each Route Domain) - associated with the single HAgroup.(No autofailback)

 

in order to instigate failover should the BGP peering with the upstream R1 and R2 go down, my thoughts were as follows:

 

create 4 pools: BGP-rd1.1 BGP-rd2.1 (R1 BGP Peers)

 

BGP-rd1.2 BGP-rd2.2 (R2 BGP Peers)

 

For F51 - Add monitor TCP 179 on BGP-rd1.1 BGP-rd2.1 and add to HA group (which already monitors the trunk)

 

For F52 - Add monitor TCP 179 on BGP-rd1.2 BGP-rd2.2 and add to HA group (which already monitors the trunk)

 

should any of the three HA group constructs fail on F51.. traffic fails to F52

 

As there is no failback - traffic stays on F52. Should R2 go down, traffic should fail back to F51 - due to BGP-rd1.2 and BGP-rd2.2 doing down

 

Any idea if my dastardly plan will work please ?.. or it is riddled with holes ;-)

 

Thanks!