Ask HN: What do you use to backup your data? | Hacker News Reader

Ask HN: What do you use to backup your data?

24 pointsukao40 comments

This question has been visited before, quite awhile ago. I was just wondering what people are using nowadays. For your personal computer and also for your work systems.

View on HN

On my Mac I've started using Arq: http://www.haystacksoftware.com/arq/

Before that I tried Mozy, iDrive, and a straight Time Machine backup. After multiple external HD failures and frustrating experiences with Mozy and iDrive (100% CPU for multiple weeks on end, extremely slow upload), I've been quite pleased with Arq.

On my non-Mac machines I just use s3cmd sync: http://s3tools.org/s3cmd-sync. I guess I could use JungleDisk, but why layer abstraction on something as simple as a file copy? :)

Three tier:

Off-site (everything): Backblaze

Off-site (small dataset immediate): Dropbox

On-site (everything): TimeMachine

I'm always kind of surprised at the number of people who have issues with TimeMachine. I've been using it since it came out, and have used it through two machine migrations and one failed hard drive. It's worked flawlessly every time.

Backblaze is configured to back up the same data as TimeMachine. I move important projects (current work) in to a Dropbox folder so that items sync there as well. Having this small, focused dataset means that I don't have to wait on a more general dataset backup to finish before my current work gets synced somewhere off-site.

Can't speak for work for confidentiality reasons.

Personal: I use macs - so I have time machine running to a Time Capsule all the time. On top of htat, my photos, documents, and music folders are rsync'd from my main machine through manual scripts to a modest homebuilt fileserver on a fairly regular basis - usually whenever I dump a bunch of new photos in (because in the end, the photos are the really valuable part, and the rest just goes along conveniently) On top of that, about once a month I plug in an external drive and the same scripts back up the same stuff to a single drive, and keep it elsewhere.

THe laptop, which I use more often, but try not to keep critical data on for any longer than I have to: - Photos get moved to the main photo library on the main mac desktop (where normal backup takes over) - A bootable clone of the laptop using SuperDuper to a pocket drive. Usually ever few weekends, at least once a month - also triggerred if I dump some new valuable data on to the machine and/or am planning to put the machine at higher risk (eg: travel). Pocket drive goes offsite. - Also, with about the same frequency, a second superduper clone to a container file on a storage drive at home.

Rationale: backups are good. The most important thing I have are my photos - they are priceless and irreplaceable. Everything else I could live with out.

- While I keep multiple copies, I need to work out some kind of better offsite solution as well.

- Email, contacts, calendar, etc. live in Google apps.

- Photos (raw files), Music, old docs live on my Mac Mini "server" on an external drive, and get backed up via Time Machine to a 2nd external drive (to guard against single drive failure). I also use Backblaze as an off-site backup on my laptop + mini.

I was using Mozy as offsite backup, but since they recently changed their pricing, the cost went from $50/year to about $800/year so I switched to backblaze. We'll see how long they last at their lower prices...

How do you backup the stuff in Google apps? If eg. Google loses your data.

For work I run Windows 7, but do my work in an Ubuntu virtualbox. That file I backup locally, to a connected USB drive, and to a NAS. The OS I don't backup at all since everything I need data-wise is on a corporate server.

For my personal stuff, I have a mixture of a large media library, and my wife's business files. The entire system gets backed up nightly (via rsync) to a primary terabyte USB drive. And then I back it up again to a different secondary terabyte USB drive. The business files (her accounting data, website source code, etc.) are additionally backed up to an offsite cloud backup service.

Every year I buy a new terabyte drive and make it the new primary backup drive. The old primary backup drive becomes the new secondary backup drive. The old secondary backup drive gets DBAN'd and sold on ebay. That way my primary backup drive hardware is never more than a year old, and secondary never more than 2.

My media library backups run the risk of a local disaster; fire, etc. But I can afford to lose all that and just be sad about it, but no big financial loss. My wife's business stuff is backed up locally + offsite for obvious reasons.

I use Git to keep track of all my code. It lives on either Github, Dropbox, or AeroFS.

Movies, Music, eBooks live on a WD World Book and aren't backed up. (no real need since I can download them all again)

The few servers that I managed are backed up by http://sourceforge.net/projects/automysqlbackup/ and pushed to an s3 bucket. Code is always in a Git repo so no need. Images and other data from user uploads are backed up by http://s3tools.org/s3cmd

http://www.dokuwiki.org/dokuwiki is used to keep track of changes I make to my.conf, http.conf, php.ini, etc. Backup is done by s3cmd again.

Images of my son are saved using AeroFS since there's no size limit.

Business documents are stored using Dropbox and Google Docs (nothing too confidential)

For documents that are pretty important, they are stored off line on a printed piece of paper and stored in a filing cabinet. IRS documents, Business contracts, etc.

All of my documents (presentations, Word documents, etc) sit on Dropbox. Most of my photos are also on Dropbox, Facebook or Picasa. I have some music files on Amazon Cloud Drive (although a lot of my files are WMAs...). Some code is on Github; but they are also all on Dropbox.

Otherwise I do a full-hard drive back up onto an external hard drive about once a week, so it makes an entire image.

For my laptop (OSX) I use CrashPlan plus[1] for minute to minute backups of my entire home folder where all my data lives. I also use Dropbox but that's more for collaboration than backup. I occasionally boot with up with my OSX CD and use Disk Utility to do a complete disk clone of my solid state laptop drive onto the original factory 2.5" DeathStar drive. My projects live in git and are pushed to GitHub and Email and contacts live at google and sync to my iPhone via their active sync implementation.

For my windows machines I used to use Mozy but it does not stay out of the way as well as CrashPlan so if I cared at all about what was on my windows computers anymore I would probably use CrashPlan there as well.

[1] http://www.crashplan.com/consumer/crashplan-plus.html

I have BackupPC running on Linux on a headless PC in my laundry room. It does daily incremental backups (file-level dedup) of itself and my laptop over ssh. The important part, as with most systems, is that after the setup process it works with no intervention on my part.

Err, hello clone? Self from alternate dimension? Don't forget to tell them that it runs on top of a software RAID5 and that pooling and compression performance is so excellent that we're stuffing about 2T of data onto 750G of disk. Also that the dryer lint doesn't seem to have any kind of detrimental effect on the computers.

BTW, don't forget that the wife wants the laundry done tonight.

Like many other folks I am on a Mac, so I have a Time Machine with 1 TB for local backup and I use CrashPlan family plan purchased around XMas when they were offering a really cool discount and got 3 years paid for. I really like CrashPlan: it took a while for the initial backup on my iMac but after a few weeks all my data have been backed up. CrashPlan is now installed on all the mac in my household (wife MacBook Pro and her business iMac). I had at one time to restore a few files for my wife's iMac and everything went smoothly so quite happy with the service.

I use DropBox too but I don't consider it as a backup solution rather a file sync solution.

I use JungleDisk on both my desktop and my personal file server. The desktop has it on both the Ubuntu and the Windows partition. I use it for the irreplaceable stuff like documents and pictures on the server, and for application data on the desktop. Laptop intentionally has zero irreplaceable data on it. I couldn't be more happy with JungleDisk despite RackSpace owning it.

The file server is 2 TB x2 RAID 1, so there's some protection there - one of the disks actually just failed, and I was able to successfully recover.

I only run hobby sites, so the databases on those I'm currently just backing up to AWS.

For contacts, email, calendar I just use the google cloud. For works in progress not on FTP or GIT, I use dropbox. I had a raid array fail on the controller, so I only do incremental backups now. I have a duplicate sized hard drive for every hard drive I need backed up. For my mac, I use superduper and for my pc / linux machines I use clonezilla which does an excellent job. The downside with clonezilla is that it is an OS so you need to have some downtime before you backup. The good side is that it backs up entire disks regardless of the operating system / partitioning / etc.

I use Duplicity (http://duplicity.nongnu.org) for incremental and full backups. It has built-in encryption as a default (but can be disabled). It also manages secure transfert to a distant host via rsync over ssh.

I use it to store my distant backups on http://rsync.net, which is most likely the best storage service I have ever used (simple, fast, reliable, relatively cheap and run by smart + nice people).

Locally I backup everything (Macs and PCs) to a 4-drive Windows Home Server which keeps data duplicated in case of a single drive failure. This also gives me the ability to stream media content to a variety of devices like my XBOX.

For off-site backups, I use the CrashPlan family plan on each of my machines.

This combination gives me the confidence that I'm covered in case of local machine failure, local drive failure and total loss of the local equipment.

I have an old DDS3 tape drive and I use AMANDA (http://amanda.org) to back up some important stuff (databases, home dirs, config files). Since it's old technology, I've never had a problem finding a replacement tape drive or new tapes.

I also mirror a few of the most important things such as photos and git repos between different computers, just so I have stuff taken care of in case of flood/fire/tape failure.

Additionally, I am looking for something that I can setup and forget. I am helping a friend of mine, he owns a retail store, setup a backup system. They do daily backups to usb drives to go offsite. I just got them a nas box, and I want to setup something that would backup all 3 of their computers main files into the nas. Just for extra layer of protection. So I am looking for something that they would not need to touch, totally automatic.

Desktop OS : Windows XP. Casper to clone the C drive (OS). xxcopy to make a backup of D drive (data). svn repository on the desktop PC.

Some apps such as OneNote sync themselves between my laptop (Windows 7) and desktop PC (XP).

For others, such as Firefox Scrapbook, I run a manual rsync (on cygwin) running via an ssh tunnel. svn and remote desktop access, too, are via the ssh tunnel.

No DropBox, no Microsoft SkyDrive.

I used to use carbonite (http://carbonite.com) but that does not back up external drives. I've since switched to Humyo (http://humyo.com)of external storage I need backed up to the cloud. A nice feature of using cloud storage is I can access all of my data anywhere I have an internet connection.

For Linux backups, you might be interested in this article, if you haven't already read it:

http://www.mikerubel.org/computers/rsync_snapshots/

The combination of cron, rsync and hardlinks, makes for incremental backups with versioning and minimal storage space, all over ssh.

On Mac at home, SuperDuper for bootable backup, plus TimeMachine, to a 2TB dual disk external with drives mirrored via RAID.

SuperDuper is great. Having a bootable backup is awesome when things go wrong. The RAID gives me some peace since I can't afford to lose this data. Backing up the backup online somewhere is my next step, I think.

I use a cron job that wakes my machine in the night and starts TimeMachine.

Independent of the tm backup, I also mirror my disk with carbon copy cloner every night.

Carbon Copy Cloner was one of the main reasons why I stayed with macs. A great software that is perfectly supplemented with Time Machine, IMHO.

My MBP has a 120 GB solid state drive and a 1 TB media drive. I partitioned 120GB from the media drive to do Carbon Copy clones so that it has self-contained redundancy.

I do incremental backups on demand with Time Machine.

tarsnap for important stuff, local external disk for my entire machine (for fast recovery). I keep a copy of my tarsnap key in a different city. My local external disk is de-duplicated with ddar (I wrote ddar).

A bunch of scripts run before the tarsnap to pull in external data such as offlineimap for all my IMAP accounts (including Gmail) and a database dump from my blog.

* Time Machine * Carbon Copy Cloner (Weekly) * Dropbox * MobileMe to keep configs, settings, address book, et al. in sync.

Time Machine and DropBox for personal stuff. Servers, remote hard drives, and DropBox for business stuff.

rsnapshot. It uses rsync to be fast. It creates directories with full snapshots of my backed-up files and uses hard-links to avoid duplicating files that haven't changed.

Nothing. I live dangerously.

Mondo rescue and LTO3 tapes

ext. drive for normal stuff. Dropbox for more important stuff

Dropbox + Backblaze.

Carbonite

Dirvish

Dirvish is really nice (if you know how to fire up a text-editor)

I run dirvish at my backup server and it fetches the data from other servers over ssh. So I have a central place to configure and monitor my backups.

(I guess that everyone at one point was about to restore some backup, only to find out that this specific server stopped pushing it's nightly backups ... somewhere last year)

That's how I do it too. It's really neat.

deja-dup.

Deja-dup too. To both Amazon s3 and to a terabyte USB HDD.

All data for a finished project goes to /dev/null and as tarball on (several) DVDs.

Was just looking into Deja Dup yesterday as a better alternative to copy/paste. If it's good enough for you two...

Multiple backup locations? Interesting- can I ask how you are handling that from deja-dup?

Reply on news.ycombinator.com