Is it possible to use an F5 load balancer to front-end an HCO high-availability pool? We are planning a 2 node setup with one server in different data centers. The virtual hostname DNS setup will not work in our environment.
Yes, it is doable. Have the F5 GTM handle the routing between datacenters then the F5 LTM handle any routing in each datacenter.
When setting up the HA pool and it asks about what DNS Type you want, Choose 'Other' and then enter a bogus DNS Server IP and Zone. You can ignore the yellow banner about setting up an Alert. The Alert is specific to using a custom script to handle the DNS updates. Since F5 will handle it in your case, it can safely be ignored.
Having SQL Always-On is key with this type of setup. You need to make sure that you have a floating MSSQL Listener between datacenters (Basic AG - Active/Passive).
Thank you @chad.every This is what I was looking for. The Solarwinds documented procedure for using DNS to handle HA failover would not pass scrutiny with my organization.
what is the error checking set in F5 LTM when the main server is STANDBY (Pool A) or when STANDBY redirect connections to Additional Web Server ? TIA
If you're talking about a F5 pool member health check, i'd setup TCP 17774 as it will only be listening on the active server.
Thanks Chad. Will single sign on (SSO) work with this setup in HA either Active or Standby (Pool A) or will only work on additional web servers (AWS) ?
Circling back on this comment. Starting in 2025.1 we added the ability to disable the SolarWinds software from managing the VIP/Virtual Hostname (configured when editing the HA pool). This allows customers to bring their own load balancer to the table without the need to follow the previous steps I mentioned of entering a fake DNS entry.SolarWinds has also publish this generic document when working with Load Balancers in the platform.https://documentation.solarwinds.com/en/success_center/orionplatform/content/core-load-balancer-requirements.htm
