March 11, 2013
Splitting a large site collection into many can make backups, storage scaling, and even development easier. This is a pattern I like to call the DLSC or "Document Library Site Collection.”
A site collection starts our small at first, gets popular, explodes in size seemingly overnight, and the SharePoint admins scramble to support it. Sound familiar?
It can be helpful to peel out a large Document Library into a dedicated site collection. Then you have more options via PowerShell and SQL to manage backups. Have five document libraries with 100GB each? Fine, split them out to 5 site collections in 5 SQL databases. Very manageable.
How To – Split a Document Library into a new Site Collection
1) Make a new SQL content database (optional). If you know LOTS of content will be coming, then give it room to grow now.
2) Make a new blank site collection. The below PowerShell command can help direct the site creation to a new database. I like to use STS#1 for “Blank Site” because it has a minimal simple footprint. Less is more.
New-SPSite http://sharepoint.com/sites/doclib -OwnerAlias “DOMAIN\Admin” -ContentDatabase WSS_Content_DocLib -Name “Document Library Site Collection” -Description “DLSC” -Template “STS#1″
3) Export the source Document Library to CMP format
http://spdeploymentwizard.codeplex.com/ can be used for a friendly safe GUI to export with all the right options. I like to enable checkboxes for ALL security, ALL versions, and ALL user info.
4) Import the source Document Library to CMP format
http://spdeploymentwizard.codeplex.com/ can be used for the import too.
5) Delete the original Document Library.
<a href="/sites/appsite/">Click here to return to AppSite</a>
Downsides and Caveats
- Manage two sets of security. Yes, you’ll have to grant people permissions in two places. Even using SPGroups won’t help because those are scoped to one site collection. AD groups can help. Those would be a great way to grant security to many site collections at once. All about scale. With only 2 sites AD groups probably aren’t worth the trouble, but with 20 it sure would help.
- Sandbox code solutions. These only run in 1 site collection, so if you’re using them to manage documents and move those document libraries you could see issues. The workaround would probably be to upgrade the code to a Central Admin farm WSP solution.
- Dependency. For some migrations you’ll need to copy ALL site collections. If you copy just 1 then you may have broken links or missing dependencies.
Hope this helps somebody else out there.