SPC 2012: Dan Holme’s ‘Deployment Wizard: SharePoint 2013 Installation Tips, Tricks, and Scripts’

SPC 2012: Dan Holme’s ‘Deployment Wizard: SharePoint 2013 Installation Tips, Tricks, and Scripts’

by | Nov 12, 2012 | SharePoint 2013

Following a quick introduction of himself, Microsoft SharePoint MVP Dan Holme began his first SPC 2012 session by explaining to the capacity crowd that “my job is to help you be successful with SharePoint.”

Part of helping you be successful includes making the slide decks for all four of his SPC 2012 sessions available to attendees sooner than later, and three of the four are already available at: http://tiny.cc/danholmespc2012, so you can “read ahead” if you’re the sort of person who likes to skip to the last chapter of a mystery novel. You see, those three decks represent Dan’s upcoming SPC sessions, but today’s session on SharePoint 2013 deployment is the one deck that Dan hasn’t yet (at the time of this blogging) uploaded to his SkyDrive, so please don’t hesitate to check my work for accuracy against Dan’s published deck when it becomes available (which should be soon).

By way of an agenda, Dan explained that he would begin with a “best practices look at architecture,” continue with a look at hardware requirements, and, finally, wrap up with a survey of the deployment options available for SharePoint 2013.

Dan stated up front that “the good news” is that deployment on the whole “hasn’t changed that much” from SharePoint 2010.

Reviewing the various server roles, Dan named the Web, Application, and Database tiers. In SharePoint 2013, these tiers can all still exist on a single server but, as Dan noted, “obviously that’s going to have some constraints on performance.”

The minimal requirement for a fault-tolerant farm is two SharePoint servers, Dan said, and “almost everything can be made redundant with this design” (just as is the case in SharePoint 2010), which also incorporates the SQL tier.

Architectural elements include load balancer; request management; Web tier (“to scale, just keep adding more servers”); distributed cache (“to scale … you can create dedicated servers”); app tier (“scale-out by adding additional servers and potentially by dedicating workloads in your environment”); query component (scale using dedicated servers with indexing), about which Dan stressed that it’s vitally important that you “start dedicating your query components soon”; and the database tier (“scale-out based on the workloads of databases”).

Regarding the decision to go physical versus virtual with your servers, Dan said that ultimately it boils down to just two things: performance and manageability.  In a nutshell, Dan said that “with architecture, there is no best practice,” and that it’s always about compromises. As a result, his advice is that you simply look at your business requirements and your server options, then “do the math and come up with your equation.”

Dan then went on to outline the minimum hardware requirements for deploying SharePoint 2013, covering both single and multi-server farms. I didn’t capture them, but you’ll find them (along with software requirements as an added bonus) in this TechNet article.

Moving on to service accounts, Dan stressed that making “sure you have the right server accounts in your environment” should be one of the most important preparatory items on your deployment checklist.  Of them, Dan said that the recommendations for the most important service accounts to have are identical to what they were in SharePoint 2010.

Addressing the notion of what Dan calls an “uber account,” he said, “I wish this wasn’t necessary, but it is.”  What Dan means by an uber account is one that will allow you to be an administrator on both the SQL and SharePoint side.

Dan said that “each farm needs its own set of accounts,” so you’ll want to take this into consideration in your naming conventions for the accounts (i.e., accounts for development, test, QA, and production).

Moving on to deployment options, Dan began with SQL, a process that (in brief) involves logging on with an SQL_Admin account, installing .NET Framework 3.5.1, and after that, running SETUP.EXE, and clicking Install SQL Features.  You’ll need to select the core features (see Dan’s slide deck), and any additional features that are necessary for your own environment.  At this point, you’ll assign SQL Service accounts.  Dan noted that, for single-server deployments, the default is for Virtual Accounts (Managed Service Accounts if virtual accounts aren’t possible) in SharePoint 2013.

Next up was the setup configuration file, and here Dan walked through a script, highlighting the four necessary modifications.  Once the modifications have been made, just run SETUP using this answer file.  (This script, as with all of the scripts he shared, will be made available at http://tiny.cc/danholmespc2012.)  Dan noted that at this point, there are just a couple of final changes to the SQL environment that are necessary.

A “little but very important note” that Dan shared is that SharePoint now requires (but doesn’t turn on by default) Max Degrees of Parallelism (MAXDOP).  Dan said that the default is 0, and though it “should be” set to 1 in 2010, it’s required to be set to 1 in 2013.  A PowerShell script to do this is helpfully included in Dan’s deck.

Next, on SQL Server, you’ll need to create a SQL login for a SP_Admin account, at which point it’s time to install SharePoint.  Here Dan said that they’re “the same four broad steps as in 2010,” and as in 2010, they can all be automated.  Dan provided a “pseudo-demo” (using screenshots) of these steps.

Regarding the Prerequisites Installer, Dan said that there are some problems to be aware of.  The first of these is a “buggy restart,” and Dan cautions that you don’t click Finish if the install won’t complete, but hit Cancel, and then run Installer manually.  As well, Dan warned that the WCF Data Services 5.0 certificate is bad (bad as in “it doesn’t work”), so the server must have Internet access during installation to successfully install.  There are also three additional prerequisites (hotfixes) that have to be done manually, but these can be automated using Dan’s provided script.

You’ll also need to Disable Loopback, and Dan has again provided a script for your use.

Now it’s time to run SETUP.EXE for SharePoint (which “has to run with elevated credentials”), so you’ll proceed to Run Configuration Wizard.  After setup, be sure to install the latest Cumulative Updates (once they exist).

Addressing a scripted install of SharePoint Server, Dan demonstrated using config.xml (which is to say, the script that comes with SharePoint).  Be advised that you will need to change the Product Key, after which you can run setup.exe/ config pathconfig.xml.

Covering how to run the SharePoint Products Configuration Wizard from the Start menu, Dan said that here again there has been no change from 2010.  A best practice Dan shared is to use a SQL alias for the database server.  During the configuration, you’ll need to specify a farm account and choose to either keep NTLM or go with Kerberos (Dan’s recommendation is to start with NTLM and make the switch to Kerberos later).

Automate Configure the Server (psconfig) was up next, and Dan shared a script for use with single-server farms.  He’ll be adding a script to cover multi-server environments soon.

At this point, as Dan said, “you have a farm.”

Dan shared some recommended resources for PowerShell including Gary Lapointe’s blog and Todd Klindt’s blog.  Dan also recommends http://autospinstaller.codeplex.com/.

Missed any of Bamboo Nation‘s SPC 2012 coverage? Catch up here:

Get a Demo
Contact Us
Newsletter Signup