I'm faced with a problem where an iRule that was written some time ago is now creating some discomfort during decommission of the Virtual Servers.
I came to understand that any referenced pools within an iRule are being monitored as well, so if the pools within it are up and running, then the Virtual Server is considered as up and running as well.
The issue came up when decommissioning some Virtual Servers, where the pool members had been forced offline, but the Virtual Server was still online and kept responding. By investigating the VS, it had an iRule attached, referencing a specific pool (which of course is online and reachable).
TL;DR:
I understand that the behaviour is by-design, as specified in the following KB:
https://support.f5.com/csp/article/K73440174
My question is: Is there a possibility to avoid this, meaning to reference a pool within the iRule so that the VS doesn't monitor it and doesn't use such pool for availability? Maybe there's a simpler solution, but at the moment it's eluding me.
If the virtual server's to be decommissioned, I'd remove the iRule and then go from there if that's causing a problem with removing it. That won't affect the backend pool(s) being referenced besides removing a source of incoming traffic.
One thing to consider is that there might be other rules which reference other pools in a similar fashion (I'm still investigating), but I'd like to completely avoid this kind of situation, if at all possible.
Ah, okay. Yes, if you're not 100% familiar with the current configuration then you'll need to basically map it out to avoid any unexpected behavior. I've diagramed our F5 layouts (vservers <=> pools) using Visio.
