Forum Discussion

DarioGB_339840's avatar
DarioGB_339840
Icon for Altostratus rankAltostratus
Nov 07, 2018

Configuring Complex Load Balancing Scenario

Hello.

 

I'm trying to configure a virtual server using an especial load balancing scenario which is driving me crazy.

 

Specification:

 

  • 2 clusters with 2 nodes each one.
  • each cluster will be served as active-passive method.
  • each node in the cluster will be served as round robin.
  • when a cluster changes to active, it will keep this status although the initial active cluster change back to up status.

 

  1. each cluster will be served as active-passive method -> solved using priority groups
  2. each node in the cluster will be served as round robin -> solved using a RR method in the pool
  3. when a cluster changes to active, it will keep this status although the initial active cluster change back to up status -> PROBLEM! I would use persistence, but I have two nodes instead of one :-(

I need to configure this using just a single F5.

 

Any help will be appreciated.

 

KR, Dario

 

  • Are you saying that for example, you want cluster B to remain active even though cluster A may come back "up"?

     

    If that's the case then i would do something like assign health monitors at a pool member level and i would configure the monitors that get assigned to the members in cluster A so that manual resume is enabled, that way you can return cluster A to being active when you desire.

     

    Or i could have totally misunderstood your requirements :)

     

    • DarioGB_339840's avatar
      DarioGB_339840
      Icon for Altostratus rankAltostratus

      Actually, I was taking in consideration to use "Manual Resume", but the point here is that both members of the cluster should work as a single node.

      Imagine this scenario from the picture above.

      Cluster-A, Node 1 - Passive
      Cluster-A, Node 2 - Passive
      Cluster-B, Node 3 - Active
      Cluster-B, Node 4 - Active
      

      In case of node 4 flapping (down-up), it should be automatically re-enable to help Node 3 to serve the traffic.

      If both nodes (3 and 4) would be down at the same time, then the cluster should change from Active to passive:

      Cluster-A, Node 1 - Active
      Cluster-A, Node 2 - Active
      Cluster-B, Node 3 - Passive
      Cluster-B, Node 4 - Passive
      

      I guess I cannot get exactly this behaviour using "Manual Resume"...

      KR, Dario

  • Are you saying that for example, you want cluster B to remain active even though cluster A may come back "up"?

     

    If that's the case then i would do something like assign health monitors at a pool member level and i would configure the monitors that get assigned to the members in cluster A so that manual resume is enabled, that way you can return cluster A to being active when you desire.

     

    Or i could have totally misunderstood your requirements :)

     

    • DarioGB_339840's avatar
      DarioGB_339840
      Icon for Altostratus rankAltostratus

      Actually, I was taking in consideration to use "Manual Resume", but the point here is that both members of the cluster should work as a single node.

      Imagine this scenario from the picture above.

      Cluster-A, Node 1 - Passive
      Cluster-A, Node 2 - Passive
      Cluster-B, Node 3 - Active
      Cluster-B, Node 4 - Active
      

      In case of node 4 flapping (down-up), it should be automatically re-enable to help Node 3 to serve the traffic.

      If both nodes (3 and 4) would be down at the same time, then the cluster should change from Active to passive:

      Cluster-A, Node 1 - Active
      Cluster-A, Node 2 - Active
      Cluster-B, Node 3 - Passive
      Cluster-B, Node 4 - Passive
      

      I guess I cannot get exactly this behaviour using "Manual Resume"...

      KR, Dario