I have a couple of questions. given the below setups on a GTM/DNS:
Scenario 1: FQDN - Pool - Pool Member 1 (Enabled), Pool Member 2 (Enabled). Both Pool members are generic hosts. No health check monitor. Global Availability.
In the above scenario, will Global Availability never function, because no health monitor is checked? Will custom service ports and custom monitors be required in order to successfully fail over traffic and responses from Member 1 and Member 2?
Scenario 2: FQDN - Pool - Pool Member 1 (Enabled), Pool Member 2 (Disabled). Both Pool members are generic hosts. No health check monitor.
In the above scenario 2, if I want the F5 DNS to practically act like a traditional DNS server, would the above config suffice? Would the service port configured under the pool members have zero relevance in this case? Would Pool Member 1 be up and green at all times, regardless of the state of the server and whether it's up or down?
In my opinion building a setup without a proper monitoring in place is a risky setup.
For your scenario 1: When there is no health check assigned, the pool members will always be marked up as available for load balancing. So in your case the 1st member will always be Up and would never fall back to the second one. Yes you need proper monitoring, atleast a service check ( a simple port 80 check or whatever service that you need to monitor) be put for proper fallback.
For your scenario 2: Traditional DNS has no failure detection mechanisms.
Thanks a lot. In scenario 1, what exactly is the sole purpose of even having a service port defined at the Virtual Server level? I always found this confusing, what purpose does this setting server? To even expand on this further, I've seen service ports with (ie, 6987), and then a custom https monitor (get http). I'm curious, how would this work and what's the purpose of the service port?
For scenario 2, would Pool Member 1 always remain active? Or does it monitor the service port by default (no specific monitors). Again, I'm a bit confused on what the service ports purpose is. I thought if no monitoring is set up, the virtual server will be unavailable/offline
The service port defined is for health check. Same like bigip monitor which tells the GTM that which all virtual servers are healthy on the LTM. Similarly if you are using generic server, you need to have some monitoring in place, be it simple http or https monitor or even a small service check monitor (tcp). This is will help the GTM know if the generic server is healthy or not. You put a service check port 6987, the GTM will make a tcp test on this port, if the port is listening on the generic server, the VS is marked available. Ultimately the same will be marked up on the GTM pool level. If the port is not listening, TCP will fail, thus GTM will mark it down. But it will always keep checking on its interval set and will mark it up when the port is listening.
Whereas https monitor is a content check, after the 3WHS, you send a GET request, if the expected response is there, you mark it up. If not, its down.