We are doing break/fix on a server that will include a reboot. Server admins are asking us to confirm that all web requests will be routed to a secondary server during the change window.
Once we are complete, we will want both web members taking requests.
What's the best way to do it?
I would do it this way:
1. Enable the feature "priority group activation" in a the pools.
2. Add new nodes with a priority group higher than the rest of the nodes of the pools.
3. Disable + force offline the old servers.
4. Remove connections or persistence records that still match the old servers.
5. Start doing the maintenance tasks over the old nodes.
6. Enable the service over the old nodes.
7. Finally disable the "priority group activation" for the pools.
8. (optional) Remove the priority group value you assigned previously to each node.
The elite of DC has spoken but I just wanted to add a small note on OneConnect. It's an awesome feature but you should be aware of some things if the following conditions are true:
Then make sure that the X-Forwarded-For headers are inserted on the F5 and that the servers can read them. Otherwise you'll see some strange traffic patterns in the server logs. 🙂
More about XFF here: