I needed to use webfarm for some old production applications. It's not typically use, because I have only one server at the time used by it (i use it for redirecting purposes to single different server). The health check is passing for my server and it works fine. But after some time without any error in both logs on farm and event viewer server turns out unavailable. The servers are excac copy of each other (i copied virtual machine), so there is the same version etc.
I tried disabling provisioning for both applications and platform, changing load balance type, change on the destination machine IddleTime on all apppools to 0, so there will be no problem with non-active apps.
There is no trace of any error in event log, only entries came, when I manually bring it up. There are also no information on the farm view, where the log is displayed. For both farm (only old ones) and server (none since i disabled provisioning).
On the server side the are no errors in the event viewer correlated to that problem. Only message that I get repeatedly is gupdate issue - Service stopped.
When server is available everything works fine. The problem occurs when I use the server as well as when the routing rule is disabled (the farm is on, I only disable the url rewrite rule itself). Time varies from couple of minutes to over one hour.