1

we are currently planning to set up a 2 tier SharePoint 2013 Enterprise farm. Two tiers with

  1. a MySQL Cluster (we are not touching this, we get it as a service)
  2. with 2 Servers in Layer 1 (Server1 and Server2 as WFE + Application). I'm using AutoSPInstaller to configure the farm.

Now I am struggling how to configure these two servers as Microsoft has not provisioned a best practice which services etc to deploy to each of the servers.

I tried to configure this on my local computer to test the configuration with VMs but it has not been very effective. I'm always running into problems when exactly installing the same to both servers. Especially with the Web Applications which of course cannot be provisioned a second time (so I'm excluding them on the second server).

If I have joined the second server to the farm I cannot reach central admin from Server2 I cannot reach it but only with using Server1 IP as the address and the corresponding port. I cannot reach any of the site collections from the web application from Server2.

Anyone has a best practice guide what to provision to the Two-Tier Architecture when using 2 Servers for increased redundancy?

1

Configuration of services and web applications on SharePoint to provide redundancy are different.

SharePoint service applications should be configured, then the services started on each of the servers on which you want to run them, therefore, for example if you want to run Excel Services on both servers to provide redundancy, configure the service application settings, then simply start the service on both SharePoint servers in the farm. SharePoint will take care of the load balancing for service applications internally for you, so if one server becomes unavailable, the other server will still provide the service for users. Once the failed server comes back online, it will be automatically be added back into rotation by SharePoint.

Web applications on the other hand require that you provide load balancing services externally to SharePoint to the servers that you want the users to access. Web applications are automatically rolled out to both servers for you however, so you don't need to create the same web application on both servers in the farm. You should find when you create a new web application that the same settings end up on both servers (look in IIS Manager). Note that you'll also need to configure DNS for the web application URLs to point to the load balancer IP VIP (virtual IP address) instead of one of the servers in the farm.

It sounds as though you've not provisioned central admin on more than one server in the farm. Ensure that the central admin site is available on both servers in the farm, and if not, run the wizard on the server that's not hosting it and modify the configuration to add central admin. Access to central admin out-of-the-box is via servername:port, however you can add another URL for central admin to respond on in Alternate Access Mappings in central admin for the web application, however you must also add the same host header for the site in IIS on each server in the farm that is running central admin.

Microsoft provides a couple of useful posters on service configuration recommendations depending upon whether you're after a traditional or streamlined topology:

Traditional: https://www.microsoft.com/en-us/download/details.aspx?id=30377

Streamlined: https://www.microsoft.com/en-us/download/details.aspx?id=37000

Post again if you need further information on specific configuration options.

  • Thanks Andy, thats quite some useful information. I will now try to provision the services etc. like you mentioned it. I did not know SharePoint is internally shifting the load for the service applications etc. I knew these both links (thanks anyway :)). I will get back after i tested it. Currently we have no external load balancing but it is on the todo list as well. – Dominik Novel Mar 10 '16 at 11:03
  • For network load balancing (NLB), if you don't have a dedicated appliance to do this, you can use the version provided with Windows as a basic option, but if you have a separate server that can be used as an NLB server, then I'd recommend using Application Request Routing (ARR) for NLB. It's a free add-in for IIS and can be found at iis.net/downloads/microsoft/application-request-routing – Andy Dawson Mar 10 '16 at 11:10
  • Thanks, i have tried the NLB Service. It seems to work. I will have a deeper look at it for the real System and machines. – Dominik Novel Mar 15 '16 at 9:33
0

As a practice, the decision about which service will be run where, depends upon your usage. I.e if you have millions of the documents for search then you need dedicated search server, same with workflow manager etc. Also, you have to decide which service you want to configure /use? with AutoSPinstaller it will install everything and configure, so be careful.

As you mentioned 2 servers( 1 App and 1 WFE). for small usage, I would configure the Search, UPA, MMS, Central admin on the App server and for customer's web application I will use the WFE.

Now, if you have FQDN(DNS) for the web application, that should point to the WFE otherwise, you will not get it.

Accessing the Central admin from 2nd server should not be an issue but some time loopback check or IE policy block it. So check that as well.

For redundancy, Web Application auto provisioned on all server in the farm, if WFE goes down, you have to update the FQDN and point to new server...or if you are using the port number & server then You have to update the AAM with 2nd server.

As mentioned in other replies, Services are auto load balanced...What you have to do just start the services from the Central admin on 2nd server.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.