Forum Discussion

garypayton_1346's avatar
garypayton_1346
Icon for Nimbostratus rankNimbostratus
Oct 16, 2013

Traffic Groups and HA

What is the tie in between traffic groups and high availability features such as vlan fail-safe or HA groups?

 

I have a new LTM pair deployment that is going to service two data centers. If there is an issue in DC1, I don't want to failover traffic that is present in only DC2. Is there a way to do this selective failover how I want?

 

As I'm reading more of the "traffic groups" chapter of the redundant systems guide, it appears that my thoughts on how this worked may be off and I can't accomplish my above stated goal...

 

availability
deployment
design

5 Replies

Sort By
  • What_Lies_Bene1's avatar
    What_Lies_Bene1
    Icon for Cirrostratus rankCirrostratus
    Oct 16, 2013

    HA features will result in Traffic Groups failing between devices. Based on your brief statements I think you should implement Active/Active where one traffic group for DC1 is active at DC1 and another is active for DC2. Let me know if you want more detail around that.

     

  • garypayton_1346's avatar
    garypayton_1346
    Icon for Nimbostratus rankNimbostratus
    Oct 16, 2013

    Both devices are connected to both data centers. I want one BIG-IP to remain active for both unless there is a problem. Really the only problem that I am trying to overcome is if the locally connected switch stops functioning properly. With that said, I think what I need to do is configure gateway fail-safe on each device, with the real switch IP address for each VLAN (ie. not the vrrp virtual gateway). If that stops responding, it's safe to say communications on that switch will not be functioning.

     

    Let's see if I am understanding this right...

     

    VLAN 1 is in DC1 and is in Traffic Group 1. VLAN2 is in DC2 and is in Traffic Group 2.

     

    In DC1, SW1 is 1.1.1.1, SW2 is 1.1.1.2 and the virtual is 1.1.1.3. On F51, which connects to SW1, I will have gateway failsafe on 1.1.1.1. On F52, which connects to SW2, I will have gateway failsafe on 1.1.1.2. F51 is active, loses connectivity to 1.1.1.1, and fails over traffic group 1 ONLY to F52. What happens to VLAN 2 and traffic group 2? Does the F5 know to only failover traffic group 1 since that is the traffic group VLAN1 resides in?

     

    Hopefully that is not too confusing...

     

    Thanks for all the help.

     

  • What_Lies_Bene1's avatar
    What_Lies_Bene1
    Icon for Cirrostratus rankCirrostratus
    Oct 17, 2013

    If you only require Active/Standby ignore my earlier response which was based on "I don't want to failover traffic that is present in only DC2". A single traffic group will suffice based on the additional details you've provided. Regarding gateway failsafe I believe you need to use this in conjunction with Fast Failover (HA-Groups).

     

  • garypayton_1346's avatar
    garypayton_1346
    Icon for Nimbostratus rankNimbostratus
    Oct 17, 2013

    This is technically Active/Active, yes. In most instances, it will be running Active/Standby but I want to react to network events to trigger a failover for one traffic group (ie. one data center), not both. If that occurs, it will be in an Active/Active state with one BIG-IP active for one traffic group and the previously standby BIG-IP active for the second traffic group.

     

    That's my goal, anyways but I'm posting here to determine if that's even possible.