Reclaiming a TimeMachine Volume’s Disk Space — 5 Comments

  1. Thanks for these tips.

    You can delete multiple backups at once using wildcards by escaping the spaces in the path, instead of enclosing a single literal path in single quotes. The following deletes a month’s worth with one command.

    sudo tmutil delete /Volumes/Time\ Machine\ Backups/Backups.backupdb/thor/2012-07*
  2. I was told (by folks at Apple’s genius bar) that it’s better practice to split up your drive into partitions so the backups aren’t having a tug of war. Do you think there’s anything wrong with just deleting the entire time machine backup of the first machine; live partitioning the drive; and then letting the second machine create a new time machine backup? I’m going through this process now, and would like to get other’s perspectives first…

  3. Since writing this post, I have had limited success deleting old backups to free up space. And I have recently come to the same conclusion as Apple advises (although I have not implemented it yet). As you put it, the “tug of war” has become an issue because 3 machines sharing a 500GB TimeMachine volume have managed to consume it all, but not evenly. (I didn’t expect they would.) That being said, I find myself running around to all 3 machines deleting old backups.

    So I think one volume per machine makes some sense, as long as each is large enough for the assigned machine. My iMac has a 1TB drive, with about 1/3 of it in use. But with TimeMachine exclusions my full backup size is just under 100GB. The other Macs in the house have smaller drives. But I am thinking perhaps a 150GNB or 200GB volume per Mac is a workable solution.

    In my case, my TimeMachine volume is actually a network volume served up by Netatalk running on my Linux server. Netatalk “fixes” the size at 500GB for the benefit of clients. But in fact, the TimeMachine volume lives on a 1.5TB RAID5 array with plenty of available space. So I can easily define new TimeMachine volumes (they get advertised as such for clients) in new directories on the RAID5 array, point the Macs at their new Time Machine volumes, and things should start working better. I would also expect to get a little better performance as well, since each client would be accessing it’s own volume rather than a shared volume.

    I cannot speak to re-partitioning an actual TimeCapsule volume, as I don’t have one of those. Sorry, but good luck!

  4. When I ran netatalk, I would create a X00 gig file, then mount it as a loopback device. Let time machine fill it up. It worked as it’s own partition, but since it was a file mounted as a loopback, no need to mess with actual partition tables. If I needed more space, create a X000 gig file, move the sparse bundle over to it and mount that one instead.

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: