Forum Discussion

  • you can find the information covered well here: https://support.f5.com/kb/en-us/products/big-ip_gtm/manuals/product/gtm-concepts-11-3-0/4.html

     

    Here is how I use them: I have three datacenters, three gtm(one in each DC), and three prober pools(one for each data center). The prober pools include all three GTM, in each of the dcs, and are set to use Global availability. In the Prober Pool, the Local data center GTM is listed as the first prober.

     

    Primarily, this was done to have deterministic probing. I don't like not knowing where a monitor might be sourced from. So when all is healthy, the objects that need to be monitored by the GTM, are monitored by the GTM within that data center. Until it fails, and one of the other GTM will pick it up. This creates prober redundancy, and maintains the deterministic nature of the architecture. This is just one scenario,.. There are many ways to use Prober Pools.

     

    Hope it helps! Good Luck!

     

    David

     

  • Salim_83682's avatar
    Salim_83682
    Historic F5 Account
    I suggest you read through this GTM doc : https://support.f5.com/kb/en-us/products/big-ip_gtm/manuals/product/gtm-concepts-11-5-0/4.html In a nutshell, they are pool of F5 devices that will monitor VS for availability. You decide who probes what and where if you have preference for specific DC or devices (called Servers in GTM configuration). Salim
  • Is there another big-ip platform in that DC that will be talking to the GTM devices, like an LTM? You can put those in prober pools too.

     

    David

     

  • I understand you have a Generic host in the DC. Do you not have any other big-ip devices in that DC? Those would what monitors the Generic host.

     

    If not,.. you will still need monitoring from the GTM devices or some bigip platform, that reports to the GTMs. So, a Prober pool would still help you be deterministic in the monitoring of that Generic host, even if coming from a remote dc. That way you have control over which GTM is primary for monitoring.

     

    I think if you get a good understanding of prober pools, you'll be able to find a scenario that works for your environment. The thing to remember is that if you don't define this, the GTMs will automagically. And that can be hard to own operationally, as the monitors can move, and the GTM may not have enough insight to know which device is really best suited to monitor the generic host, so could assign it to one that cannot reach the generic host, even if it is reachable by the client.

     

    Good luck!

     

    David

     

  • Understood David, there is no LTM platform in the DC's this new Generic host will be configured. Out of 2 DC's with GTM's, one DC is primary and the other is Dr only. I would want DR GTM monitor this generic host only if GTM in primary DC fail.

     

    I use Layer 7 HTTP monitor to check the health of the VS on the server, and I think with that monitor GTM would check the availability of the service externally so if the primary DC cannot reach the url then Primary GTM should fail the traffic to DR GTM (provided it could reach the url externally).

     

    So since both GTM's are in remote DC's, do I really need a prober at all? Thanks

     

  • If you are running a monitor from a GTM, that IS the Prober. The GTM SYNC group will determine which is closer, and use it. Unless you build a Prober pool. If you define a prober pool, assign it to the DC your generic host is configured in. Configure the Prober pool to GA LBing, with the Primary GTM listed first,..... You will have what you want.

     

    David