Hacker News Reader

Comment by brucehart | Hacker News Reader

For anyone looking to an alternative method, there is a great open source Python script called Bagoma that incrementally backs up your Gmail account: http://bagoma.sourceforge.net/ . Each email is stored as an individual text file rather than the mbox format. I have backed up my account with it every few days for a couple of years and it works great. It will even restore your account to Gmail if you need to switch accounts or accidentally delete messages.

View on HN

A similar project is http://gmvault.org/

I'd never heard of bagoma before. I've used Gmvault successfully on both gmail and google apps accounts.

I've never heard of either and just tried out gmvault and it's currently running and appears to be working.

Anybody tried out both and have data to share?

An alternative to this would be fetchmail and maildrop. The MTA (fetchmail) downloads the messages periodically and delivers them to the MDA (maildrop) which can store them in Maildir format (each mail is an individual file). Fetchmail actually has weird bugs with some mails, so I made my own version in Perl, but there's other options out there.

Alternatives: http://imapsync.lamiral.info/ http://offlineimap.org/ http://isync.sourceforge.net/

Please don't use fetchmail -- it loses data. Use offlineimap or mbsync instead.

Watch out, mbsync reacts very poorly when Gmail decides to change UID schemes (UIDVALIDITY), which happens any time a label is created or removed.

fetchmail is a complete POS. Don't use it.

Citation?

IIRC, the semantics used to be that if

1. You used the keep option,

2. fetchlimit was set to be less than the number of messages in the remote mailbox, and

3. The remote mailbox stayed above this limit,

then you would find yourself in a state where fetchmail would say that it had fetched everything while there were unread messages that it would never look at.

Looking for a bug report just now, I didn't see anything that matched this.

Maybe this?:

http://www.mhonarc.org/archive/html/fetchmail-friends/2003-1...

While I can understand holding a grudge against any software that drops email -- that was in 2003...

[edit: also see markrages link on "why getmail" below]

http://pyropus.ca/software/getmail/faq.html#faq-about-why

I don't know about lost messages, but it used to bomb out on spam, but still deliver the messages up to the spam. So I would wake up with 5,000 duplicate messages and a lot that were never gotten, and i'd have to use a client to delete whatever seemed like the "bad" message until fetchmail could continue delivering mail and finally delete the old ones.

To solve it I did what any bored Perl programmer would do - write my own fetchmail program in a couple hours in Perl. It solved my issues and i've been using it ever since. Years later it took me about an hour to add SSL IMAP support so I could use it on Gmail and Exchange.

FWIW I needed to do this to back up business emails when I closed down my last shop. At the time, all the open source projects I found didn't manage to correctly download ALL messages. I was finally able to do it by doing an HTML inject, and using JS to load each email iteratively.

Did you try bagoma? When was this?

2-3 years ago, and I don't remember, unfortunately. I just wanted to point out that it could be done through simple javascript if those viewing this thread are unhappy with any of the libraries mentioned.

Another tool which does a good job is Mailstore, free for home users:

http://www.mailstore.com/en/mailstore-home.aspx

Reply on news.ycombinator.com