Beginners start here: Time Machine Exclusions

Sean McNamara
26 February, 2009
View more articles fromthe author

Time Machine has revolutionised backups on the Mac – but there are some issues around how it operates which can see backup drives fill up quickly, or data being backed up unnecessarily. This month, we’re going to look at some ways to fine tune what Time Machine will back up.

When it comes to talking about backups, I probably sound like a broken record. “You can never have too many backups.” “Backup, Backup, Backup.” The beauty of Time Machine is that it does hourly backups of anything that’s changed since the last backup, keeps the last 24 hourly backups, and keeps one a day for the last month, and then one a week before that until the backup drive is full. Now that’s a quantity of backups I like.

But there are two general areas which users may not want backed up at all or that often.

The first of those areas is System files and applications: they can take up a fair amount of room, and some users are willing to forego the “restore a whole machine” feature of Time Machine to keep more space on the backup drive for data.

Although Time Machine doesn’t have many controls, it does have the ability to exclude files or folders from backups, which allows some level of fine-tuning of backup drive use.

This is done through the Time Machine System Preferences pane. Click on Options and you’ll see a file list to which you can add files or folders, either by clicking on the + button and then navigating to the item to exclude in a file selection dialog box, or by dragging files or folders to the file list area of the window.

One of the problems with wanting to exclude System files is that some of them are invisible, and would be difficult to add to the list of excluded items.

Luckily, Apple has included smarts in Time Machine so that if you exclude certain visible System file locations, it will prompt you to see if you want to exclude all System file locations.

Another “problem” for Time Machine (from the point of view of potentially filling your backup drive faster than you’d like) is large, often changing data-sets. The sorts of things I’m talking about here include (but are not limited to) Entourage databases and VMWare or Parallels virtual machines.

Entourage stores all of its data in a monolithic database which can grow quite large. It can just grow large if you never delete your messages, especially if you receive many attachments.

However, one “feature” of its operation means it will continue to grow, even if you put messages in Entourage’s Trash and empty it – when you delete messages in this way, you’re only deleting a reference to it, not the actual message. If you think of messages as pages in a book, which has a table of contents. Obviously, as you receive more messages, there are more pages in the book (the database gets larger). However, deleting a message in Entourage is analogous to crossing entries off the table of contents but leaving the pages in place. The book (database) continues to grow, no matter how much deleting you do.

In the past, Entourage would only support databases up to a certain size (at one point it was 2GB) – if you ignored the warnings as you approached that limit, you could actually irretrievably corrupt your database when it exceeded that limit.

You can optimise the database by holding the Option key down while starting Entourage, which will effectively rebuild the database with only the messages which haven’t been deleted (like tearing the pages of those messages out of our example book), but Entourage still keeps the original database (which is a good idea, just in case something goes wrong), taking still further room up on your hard disk (and, by extension, your Time Machine backup disk).

When this ever-growing behaviour is combined with the fact that everytime you do anything with Entourage (open the program, read unread messages, reply, etc.) you are marking that monolithic database as changed, you end up with a situation where a potentially very large file is being seen by Time machine as changing very often, which will make your hourly Time Machine backups take quite some time, and your backup HD will fill very quickly.

For this reason I tend to recommend one of two courses to Entourage users if they don’t want these problems – either exclude the Microsoft User Data folder (in the Documents folder in your Home folder) from Time Machine and back it up manually (or automatically with some other backup program), or switch to Apple Mail (which stores each message individually, so only the few individual messages which have actually changed since the last back up are backed up).

Another example of a large data set which may change often enough to quickly fill your backup drive is the drive image used by virtualisation products such as Parallels or VMWare Fusion.

Simply because these drive images are often representing multiple-gigabyte hard disks, they can represent a large proportion of your backup hard disk use.

Both VMWare and Parallels have different features which can allay this problem: VMWare can use disk images which are “sliced” into chunks of a size of the user’s choosing, so only those chunks which have changed will be backed up by Time Machine (but you never know which chunks, or how many, will change for any given operation in the virtual machine), and Parallels allows you to set individual virtual Machines to not back up in Time Machine.

I tend to recommend that virtual machines not be backed up in Time Machine (unless you have a very large backup drive), and that they get backed up in some other fashion. Parallels virtual machines are stored in the Library folder (exclude the folder called, appropriately, Parallels) or use Parallels’ preferences to exclude individual virtual machines in the settings for those virtual machines. VMWare’s virtual machines are stored in the Documents folder in your Home folder, in the folder named “Virtual Machines”.

One other tip for getting data from virtual machines backed up is to store those documents in the Mac OS X filesystem and share them to the virtual machine. That way, only the documents which have changed will be backed up by Time Machine. Both products have support not only for arbitrary folder sharing between Mac OS X and the virtual machine, but also sharing the user’s Documents, Desktop, Pictures and Music folders to the PC’s My Documents, Desktop, My Pictures and My Music folders respectively.

One last tip is that if you’re temporarily working on data, create a “Time Machine Excludes” folder on your Desktop (or in your Home folder, or in your Documents folder, or anywhere you like, really) and exclude that folder in Time Machine. Now, anything you put in that folder, no matter how large or how often it changes, will not be backed up. I use such a folder to store Outlook .pst files (which, like their Entourage counterparts can get quite large) for converting via Parallels, and large MYOB files which I need to work on regularly but which are backed up on their original machine. I also place this folder on my sidebar so it’s easy to drag items to it or use the save and open dialog boxes to access it.

Don’t forget, though – anything you exclude from Time Machine which you want to back up will have to be backed up by some other means, whether manual or automatic. You should only exclude data you don’t mind losing, or which is backed up some other way.

Leave a Comment

Please keep your comments friendly on the topic.

Contact us