Hacker News new | past | comments | ask | show | jobs | submit login

Summary of the article

> It’s gone because it was kept entirely on his blog

Take backups.




They closed his email-account too, though? It has never occurred to me to backup my entire Gmail archive. Should probably look into that now...


Admittedly email has also been in my blind spot too.. Not a bad plan that.


Here is a one-click service to backup everything in your google account. Email, google drive, youtube, location history, blogger blogs, bookmarks, calendar, literally ever google service.

https://takeout.google.com/settings/takeout


Nice tool, and can produce archive directly to Dropbox (or Drive, or OneDrive).

Now, if they only could add a periodic option...


What format does this produce? A bit of googling suggests mbox. What happens to 'labels' on emails?


MBOX is correct. Labels become folders. Mail that has multiple labels will be duplicated into each folder.


If you click "details" you can see the format each service is exported as


I started archiving all my emails a long while back. There are several good tools around, but I'm currently using offlineimap[1][2].

Easily installable via homebrew.

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

[2]: https://github.com/OfflineIMAP/offlineimap


The one gap here, is that if Google closes down your email account you can still be dead in the water if you have any services that rely on that email. Such as email based 2-factor authentication, or any websites that you forgot your credentials for an need to rely on an emailed password reset.

You can mitigate this by having your own domain and host the email on Google Apps -- that way if they take away your access you just update your mx record to point to somewhere else. But you can also easily lose your domain name (due to malice or accident).

BTW, if an emergency does happen, anyone have a recommendation for a decent web email client that I can point my MX record to? Preferably one that has both hosted and self-hosted options. Or something that I can spin up easily on a cloud service such as AWS.


A couple of open source clients [1] and [2]- both self-hosted and deployable on AWS or any other cloud service, or any other hosting service with ssh root access:

[1] https://squirrelmail.org/about/ [2] https://www.horde.org/apps/webmail

As an aside, VestaCP [3] - an open source hosting admin panel that you can install within a few minutes via ssh - will install squirrelmail automatically:

[3] http://vestacp.com/


I started deleting my emails and I feel really better that this big stack of crap is erased forever. The whole concept of archiving mail is disgusting. Everybody is concerned by climate change, but seriously, storing all these data is crazy and dirty, a datacenter is everything but clean.

No backups to deal with and no procrastination-until-the-end-of-your-life. Save the planet and get away from your past.


I have two decades of email saved; it's a few gigabytes and could fit on a single USB stick or DVDW.

I happen to like my past and want to keep it, even if I don't habitually refer to it.


this looks interesting. Does it also restore to another IMAP provider?


mbsync [1] is worth looking into. I found it to perform better than Offline IMHO and complained less. It will store downloaded emails in maildir format.

[1] https://wiki.archlinux.org/index.php/Isync


Speaking of that, what is the best way to make a backup of the contents of an email account (not specifically Gmail, but IMAP in general)?


An easy way is to use an IMAP client like the macOS Mail.app that syncs raw email + metadata in .mbox format. You can find the files in ~/Library/Mail/V4. Combine with Time Machine for extra backups in case you’re worried that your mail server decides to clear your local copy over IMAP.


Not sure if it's the best way or not, but I use offlineimap...


Any hosted email account can be disappeared without recourse. Particularly a free one where the company owes you nothing.


At least the article opens with the conclusion, saves scrolling.

Having said that, I did continue scrolling and the article goes on to say

> More generally, relying on the internet as an archive is dangerous.

And it is, at least, another point toward helping me feel vindicated for being that person who always says "and you have off site backups for when your house burns down, right?"


Not that people ever listen. Backups are perhaps a lot like basic insurance in that way. In a lot of countries basic health and liability insurance policies are mandatory because a sizeable chunk of the population would refuse to have these without that requirement, even though it benefits the vast majority of people.

Silly if you consider how from a technical and practical standpoint nothing prevents you from carrying a backup of every important document you've ever owned and a good deal of photos and such with you in your wallet on an encrypted mini-SD-card. Wallet gets stolen? Make a new backup. House burns down? You have your backup. Drop off a card at friends or family, keep one at your employer, hell, stick them on the Christmas cards you send out each year to your nearest and dearest; with proper encryption you are the only one able to use them anyway.

It also doesn't help that people blindly trust the big tech companies to do this for them.


Definitely feel vindicated, it's the 3-2-1 rule

3 copies

2 different forms of media

1 offsite


Can you qualify what "2 different forms of media" means? I'm assuming one physical, one cloud. Do you consider two separate external HDD's with the data on it to fail that test because they are the same form?

When my music collection was smaller, I had everything backed up onto an external as well as had them all physically burned onto like 50 DVDs. Since my collection has grown beyond a terabyte, however, I haven't had the time or energy to keep that up. Now it's just an external, a travel external, a redundant external, and a cloud backup.


A few examples I can pull off the top of my head - DVD, Disk (HDD), Disk (SSD), Disk (Flash.. maybe?), Tape, Cloud

So Disk (HDD) and Cloud is fine. Two HDDs (spinny platter style) would indeed fail the test. If you want to argue it you could probably get away with different manufacturers or different batches but I'd prefer my backups to play nicely with the rules rather than try to avoid them! Far-fetched as it is but lets say for example you have a power surge that takes out your external disk (is that even possible?). Chances are high whatever did that would also do that to your other disk. Or maybe a wall caves in and decimates your PC (or you just sweep the desk somehow and knock everything off it. Maybe the desk has collapsed) - The spinny disks may be destroyed if they are both hit hard enough. Mixing an HDD and SSD, the SSD would probably be more hardy against shock damage.

From another comment I made in this thread somewhere "I consider "cloud" as its own media type to avoid the pedantics of me storing files on a spinning disk and the cloud provider does too. Chances are they have server grade disks and their own backups [so it's different to just a raw disk]"

Edit: I use too many brackets here. I am going to attend my next available parenthesis anonymous support group.


One nice thing about the "2 different forms of media" rule is that it would protect you against bad batches of hard drives.

Sometimes, when people buy multiple drives at once for servers, they wind up with drives from the same batch. If the drives are subtly defective, and if they're always run under identical conditions (such as in a RAID array), then I've seen multiple drives fail within days of each other.


Running two identical disks in a RAID, or where the same desk could cave in and sweep them off, seems like it should be covered by "1 offsite" rather than requiring two different types of media.


Adding my own[0] rule here:

Disks exists in two states: failed and failing.

[0]: please feel free to let me know if you know any previous references to this rule.


Does three copies include the original?


I believe it does traditionally, yes. Your data plus two backup copies. Personally I'd say no though, just for the sake of having more backups (3 2 1 is minimum, not strict!).

Partially unrelated I consider "cloud" as its own media type to avoid the pedantics of me storing files on a spinning disk and the cloud provider does too. Chances are they have server grade disks and their own backups. On that note, their backups are not my backups. Their backups are never my backups.

If they save my butt one day, awesome. But that day was already terrible if I'd managed to lose all of my backups.


> But that day was already terrible if I'd managed to lose all of my backups.

I have two off-site backups located ~150 miles from each other. If I (we) lose everything, chances are that there are more important things to worry about, like what's for dinner.


That reminds me of my first job (late 70's)where one of the project leads was paranoid about losing data from experiments (eficiny of mixing in tanks)

He had 2 copies on floppy two on disk (in a separate building) and another set at home on floppy's

That's why in 1979 I used to by 8 inch floppy's in multiples of thousands


And don't forget to test your restore.


Actually good point that should be the first rule

0. It's not a backup unless you know it restores

1. Always take backups

2. Backup the backup


A few month ago someone here on HN stated "People don't want backups. People want restores". Can't recall who, though...


I've known (and retold) a similar saying for years: "Nobody wants backup. Everyone wants restore."


I think people are jaded due to windows automated backup, where you tell it to keep only the latest backup and it still manages to fill your hard drive with 50 of them.


Does blogger make that easy / possible?


Yes there's Google Takeout, but in this case I'd more back up the originals rather than whatever comes out of Google's optimizations


Sure, if we're talking about saving a copy of an image, I agree. I assumed that there was original content (i.e. writing) that had simply been entered directly into blogger, which a typical writer might not know how to backup at all.


Thanks Dad.


Perhaps the internet archive's "way-back-machine" has a backup?


Do people even read the posted articles at all?!

From the article:

Cooper should’ve backed his work up, or known to be more cautious with drafts. ... ... Admittedly, not everything has vanished. The Internet Archive’s Wayback Machine has saved bits and pieces of Cooper’s blog dating back to January 3, 2012, but they are just snapshots of the blog’s front page for a given date, not capturing the blog in its entirety.


Probably not a backup of his email though!




Consider applying for YC's Spring batch! Applications are open till Feb 11.

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

Search: