The traditional approach to zero-downtime restarts is to spin up a secondary server and have something like Server::Starter or HAProxy in front of your servers. Then, you tell one of your servers to gracefully spin down by processing all requests that are still in flight while already telling your proxy not to send new requests to that server instance. Once the instance has gone idle, you restart it and reconnect it to the proxy.
I don't work with IIS so I don't know if there is a built-in mechanism for that, but as long as you have multiple (virutal) machines behind a proxy, you can use the technique even without server support.
In reply to Re^2: IIS Restart
by Corion
in thread IIS Restart
by trell
For: | Use: | ||
& | & | ||
< | < | ||
> | > | ||
[ | [ | ||
] | ] |