Mozilla Thunderbird and Time Machine

[Jacquesvw]

Hi everyone,

I would like to migrate from Outlook 2011 to Mozilla Thunderbird, but know that in the past Thunderbird used to store all emails in one large database. This obviously is not a good situation for Time Machine, because a single email will result in the entire database being backed up every hour. Since I currently have over 57000 emails that must be preserved the database file will be huge.

Long ago I read that the developers of Thunderbird was planning on creating a "one file per email" kind of storage system that will allow Time Machine to incrementally backup the latest email messages and not the entire database every single time.

I cannot find any info on whether this was implemented or not, so my question is: Can a Thunderbird user please check how emails are stored and let me know if it's a single database or multiple files?

[ConCat]

There's a trick I learned ages ago for dealing with this: If you symlink a large file from a sparsebundle disk image, then only 8MB of the file will change at a time. It sorta fakes block-based backups. Not every app can cope with a symlink though, but most of them can. I can help you set it up if you want to try it.

[Jacquesvw]

Thanks for the reply. I would really like to know how your trick works, because it will be useful for VMware Fusion images as well.

[ConCat]

Originally Posted by Jacquesvw

Thanks for the reply. I would really like to know how your trick works, because it will be useful for VMware Fusion images as well.

Sure. You can create a sparse bundle disk image by entering the following in Terminal:

hdiutil create -size 20G -fs JHFS+ -volname "My Volume Name" ~/SparseData.sparsebundle

Replace "20G" with however much room you think you'd need and "My Volume Name" with whatever you'd like. Realize though that the sparse bundle will not take up more room than there is data on the disk, even if it's formatted as larger, so, a 20G sparse bundle image with 100MB of data on it will only take up 100MB+normal filesystem and disk image padding. It's quite efficient that way. When you mount the disk, you can move data onto it, then in terminal type

ln -s "/Volumes/My Volume Name/MovedFile.dat" "/path/to/original/location/MovedFile.dat"

Keep in mind that you'll have to open the disk image again if you reboot your computer, but it's only a very minor inconvenience in my opinion. I use this quite a bit. You can just keep everything on the same disk image too. All of your large files. Makes things easier that way.

Of course, for things like a VMWare image, you can just specify that you want the image stored on the sparse bundle during creation. Much simpler that way.

[Jacquesvw]

Thanks for the help, was of great help to me. Appreciate it.