IIS 7 and Above
Application request routing as reverse proxy for Sharepoint servers
Last post Jun 11, 2019 10:17 AM by ironmanwannabe
Jun 10, 2019 09:28 AM|ironmanwannabe|LINK
Hi, I have some questions below.
1) I have setup an ARR server as a reverse proxy to 2 Sharepoint servers. I noticed that ARR is does not monitor the health of the content servers if I did not configure the Health test as shown below. Without health
test configured, even if i shutdown a content server, monitoring and management pane will still shows the server as available and healthy. Client may hit the offline content server and "Fail request" will be shown on the monitoring and management pane. Is
this the correct behavior as I read that ARR is able to determine the heath using active traffic on top of explicit URL testing?
2) In the server farm, what is the use of the status of the servers? Seems like it is for us to enable or disable manually. It doesn't change its state when the server is shutdown.
3) For health test configuration, I created a test html in sharepoint site but I received a 401 error. I changed extended the acceptable status code to 499 but the error became "The response from the server did not include the configured string". The health
test works if I enter the stopped "sharepoint - 80" website and start the default website. Is there any additional settings required for ARR to monitor sharepoint URL? I can access the test URL from IE within ARR.
Jun 10, 2019 04:27 PM|Rovastar|LINK
Ok your health check looks wrong. You are connecting to localhost I haven't seen that before or wouldn't do that and tbh don't understand the behavior of what it will do.
I would use the actual site name as the health check.
So if the site is www.mysite.com I would use that is the health check. It is unclear if you want to send the health check request as https or http. That depends on what if you are offloading your SSL but do the same as
you are doing in your global URLrewrite rules for sending it out this farm.
Also I use IP address rather than names of servers for servers in the farm.
I would not use the active traffic settings. It seems pointless and will likely cause more confusion/harm than good.
There are hundreds of ways of setting up ARR so it id difficult to say for every case. I think there is different behavior for example using names of server for servers in the farm can cause different behavior (maybe that is why some edit their ARR host
2) The status of the servers.
This is for servers in the farm and if they are subject to the health check or not. In "Monitoring and management" you can manually take these servers out of the farm of these selected servers.. Restarting the farm/app pool will return all servers to the
default (available) state.
They are separate as you may want to add a server to farm but make it (un)/available based on health checks.
3) check the IIS logs of your backend servers to see if the traffic is making it through. That will tell you more. You should see the traffic hitting you backend boxes.
Jun 11, 2019 10:17 AM|ironmanwannabe|LINK
Thanks for your response. How to insert pictures into this message?
1) I tried using the servername of the sharepoint server in the health test but the health test shows the same result as using localhost. I did not configure alternate access mapping so the actual site name is the hostname of the sharepoint server itself.
My sharepoint is running on HTTP only. I am using hostname as my ARR server is joined to the same domain as my sharepoint servers.
2) How do you mean restarting the farm/app pool? Do you right click on the farm and "take server farm offline" and online? If I set a server as offline manually, recycling the farm does not turn the server back online.
3) I will check the IIS logs for clues.