100 % availablity using f5 to two IIS servers configured for load balancing in round robbin

silaspeter
silaspeter used Ask the Experts™
on
is it possible to get 100% fail over 100 % availablity using f5 to two IIS servers configured for load balancing in round robbin.

Each request to check if a server is active and route the call to the other server if one is down, without any downtime.

Also is it possible to queue a call if both servers are down and wait a certain period of time and make the call to the first available server?
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
David FavorFractional CTO
Distinguished Expert 2018

Commented:
100% uptime can be approached + tough to cover every contingency.

I did this type of work at IBM for 10 years + on my own for another 12 years since leaving IBM.

No matter how much you plan, outages will likely occur.

Start your planning by considering post-outage recovery, so how system recovers if all nodes are offline + then one comes back online.

Starting from this point will likely give you a far more robust system, then trying to design from the standpoint of one node always online.

Also, I'd suggest you hire someone to assist you with this.

IBM use to charge $250K (in 2005) just for on paper design of this type of system.

If there's huge money at stack, where outages are best avoided, then hire someone to assist you.

Also, your next investment will be someone to track + manage this type of system.

The common direction from IBM to clients, was simple monthly testing.

Have someone talk into a data center + pull the power cord out one primary/master machine, to ensure fallovers always work.

You only have an HA (High Availability) System if you continually test it + it works.
Aaron TomoskyDirector of Solutions Consulting

Commented:
Consider either doing IaaS in azure or even PaaS like IIS as a service.
https://docs.microsoft.com/en-us/azure/app-service-web/web-sites-migration-from-iis-server

Author

Commented:
We already set this up, my systems team set it to 15 seconds timeout, every call that hits one of the two servers on which an application is , if the application is down (app pool is stopped) would keep failing within that 15 second window. Each application is bound to a unique port. If i turn the app pool off, the call fails for 15 seconds maximum and then f5 knows it is down and marks it as down. It then routes the subsequent requests to the other server. (within the next 15 seconds if F5 sees the server being up it then routes the calls to that server again)

I am asking the team if we can check the server status real time on that port, and then route the call to that server only if it is active else route to the other server (if it is active).

To take it further i was asking if the requests can be queued for a certain period of time if both the servers are down. It keeps trying within that window of time which is pre configured and fails only after that if neither of the servers are active.

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial