Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: How do you backup your email?
30 points by kaptain on April 12, 2020 | hide | past | favorite | 17 comments
I'm looking for a solution for email backup. Ideally, it would not run on my desktop, it would be server-based solution that connects to my IMAP-based email provider, and saves backup snapshots to cloud-based storage (e.g. Backblaze, S3, etc.) using de-duplication to save space.

> ... server-based solution that connects to my IMAP-based email provider, and saves backup snapshots to cloud-based storage (e.g. Backblaze, S3, etc.) using de-duplication to save space.

You could rather quickly and easily write a shell script to:

- use offlineimap [0] or imapsync [1] to perform a local sync of your mailbox to local mbox files or maildir directories,

- (optionally) use tar or similar to create a single file out of your entire mailbox, and

- use tarsnap [1] to back it up (to AWS) or s3cmd [2] or similar to back it up to wherever you want it.

I used to do exactly that with an additional step to encrypt the tarball (with GPG) before uploading it. It worked out quite well.


[0]: http://www.offlineimap.org

[1]: https://imapsync.lamiral.info

[2]: https://www.tarsnap.com

[3]: https://s3tools.org/s3cmd

Note that if you're using tarsnap there's no need to create a tarball from your mailbox (tarsnap does that for you) and there's no point encrypting your data with GPG (tarsnap encrypts your data for you too).

If you're using s3cmd to back up your mail then the "create a tarball" and "encrypt with GPG" steps are useful, of course.

Some email gets printed out, stuck in various binders, and kept for awhile. Then usually set on fire in a ritualised celebration of the impermanence of reality. This is also known as scheduled secure destruction.

The majority of serious email get copied (IMAP or for the email server we run, just sftp) to a pair of DVDs on periodic basis as an encrypted filesystem archive. We then stick it in a dvd swapper drive for offline access. The other copy gets put off site.

To expand a little:

Just a computer with a LOT of dvd drives. Add a few high capacity local ssd, harddisk and plenty of ram. Insert disk, optional full copy to local as cache, then server just reads. If it copied it auto-ejects.

Also got a tapes as a backup. Works similarly. Insert tape, it gets copied to the libraries local storage then ejected. We can also just use the tape without copying first.

Why optical for email? We can easily store them in a human accessible way. They're supermarket cheap. Quicker than tape as well. Everything on optical is also on tape with no exclusions.

Its a pity optical media hasn't reached the point where I can just go buy a pack of 1TB disks from the supermarket. I'm not holding my breath for Sony's Archival disks either.

Dvd swapper drive? Does it work like a tape library? Do you have any pointers where to find these?

Commercial solutions can hold 60-300 disks or more. Expensive.

For home use just get a bunch of usb dvd drives and go nuts, or go full tape. Once you go tape you get addicted to capacity. But it gets pricey. Cost/density is interesting.

At home I just use a stripped down version. Literally a bunch of usb dvd drives hooked up to a raspi. I can insert a disk and poke around or have it auto copy the whole thing and auto-eject. I have 10 slim drives connected... ebay is kind. The raspi pipes to a different server full of ssd/hd/ram depending on what I'm doing.

Gmail: download via takeout, keep it local, backblaze to backup from local to other server

I just did a write-up on using imapgrab.py :)


Hrm. This is very interesting. Your post made me think about using a couple of different software packages together. I could use IMAP Grab to pull all the email to disk on the server and then use https://www.duplicati.com/ to push to the cloud storage of my choice. Sigh. Sounds like another project for me. Thanks! Your post was super helpful to get me thinking about how things might work.

great tip, thank you - just set this up, hauling down backups of email from over 1.5x decades from 9x email accounts into mbox !! woot ;-) reply

I use getmail[0] on my rpi. Download everything from gmail and fastmail using IMAP (sync) overnight into a truecrypt container, and then offsite backup to rsync.net. It's also storing everything into an mbox so you can easily import that file anywhere.

[0] http://pyropus.ca/software/getmail/

I use gmvault to replicate my personal gmail to my local NAS server on a nightly basis. I also have offsite encrypted backups using restic, also done nightly.

I wish I could back up to PUNCHED CARD still ;-)

Digital to paper backup:


check out http://metallic.io/ ASAP !!

they have an awesome free trial running right now..

POP3 client

isync and cron

Applications are open for YC Winter 2022

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact