Slashdot Mirror
Online Website Backup Options?
pdcull writes "I can't be the only person on the planet who has this problem: I have a couple of websites, with around 2 GB of space in use on my hosting provider, plus a few MySQL databases. I need to keep up-to-date backups, as my host provides only a minimal backup function. However, with a Net connection that only gets to 150 Kbps on a good day, there is no way I can guarantee a decent backup on my home PC using FTP. So my question is: does anybody provide an online service where I can feed them a URL, an FTP password, and some money, and they will post me DVDs with my websites on them? If such services do exist (the closest I found was a site that promised to send CDs and had a special deal for customers that had expired in June!), has anybody had experience with them which they could share? Any recommendations of services to use or to avoid?"
Rather than "posting DVDs" I'd go for something like Amazon's S3 and just dump the backup to them. Here is a list of S3 Backup solutions that would do the job.
I've personally moved away from hard-media as much as possible because the issue on restore is normally about the speed to get it back on the server and its there that online solutions really win as they have the peering arrangements to get you the bandwidth.
An Eye for an Eye will make the whole world blind - Gandhi
It seems like the only problem with your home computer is FTP. Why not use rsync, which does things much more intelligently - and with checksumming, guarantees correct data?
The first time would be slow, but after that, things would go MUCH faster. Shoot, if you set up SSH keys, you can automate the entire process.
Oh, you're not stuck, you're just unable to let go of the onion rings.
Then what you need is rdiff-backup, works like rsync except it keeps older copies stored as diffs.
As for FTP, why the hell does anyone still use ftp? It's insecure, works badly with nat (which is all too common) and really offers nothing you don't get from other protocols.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
You may have to use extra tools to break your archive into seperate chunks fitting Gmail's maximum attachment size, but I've used Gmail to backup a relative small (~20mb) website. The trick is to make one complete backup, then make incremental backups using rdiff-backup. I have this done daily with a cron job, sending the bz2'ed diff to a Gmail account. Every month, it will make a complete backup again.
And a seperate Gmail account for the backup of the mysql database.
This may be harder to do with a 2GB website, i guess, since Gmail provides atm about 6GB of space which will probably last you about 2 months. Of course you could use multiple gmail accounts or automated deletion of older archives...
But seriously, 2GB isn't too hard to do your from own PC if you only handle diffs. The first time download would take a while, but incremental backups shouldn't take too long unless your site changes drastically all the time.
rsync to get the data, cp -al to keep snapshots. I've been using this for years to manage TB of data over relatively low-speed links. You'll take a hit first-time (so kick it off at night, kill it in the morning, and the next night just execute the same command and it'll eventually catch up, then cp -al it, then lather rinse, repeat. This page: http://www.mikerubel.org/computers/rsync_snapshots/ has been about for years. Use it!
...because if you are and you're planning to sent personal data (there can't be many 2GB web sites that contain no personal data at all) on DVD through the mail, you might want to look at recent pronouncements from the Information Commissioner. A large fine could be waiting for you if you go down that route.
Then what you need is rdiff-backup, works like rsync except it keeps older copies stored as diffs.
Another option is to use the --link-dest option to rsync. You give rsync a list of the older backups (with --link-dest), and the new backup is made using hard links to the old files where they're identical.
I haven't looked at rdiff-backup, it probably provides similar functionality.
Part of my backups script (written for zsh):
setopt nullglob
older=($backups/*(/om))
unsetopt nullglob
rsync --verbose -8 --archive --recursive --link-dest=${^older[1,20]} \
user@server:/ $backups/$date/
There is also the --backup --backup-dir options (you'll need both). It keeps a copy of the files that have been deleted or changed, if you use a script to keep it in seperate directories you'll have a pretty good history of all the changes.
New things are always on the horizon
I use rsync on a few dozen systems here, some of which are over 1TB in size. Rsync works very well for me. Keep in mind that if you are rsyncing an open file such as a database, the rsync'd copy may be in an inconsistent state if changes are not fully committed as rsync passes through the file. There are a few options here for your database. First one that comes to mind is to close or commit and suspend/lock it, make a copy of it, and then unsuspend it. Then just let it back up the whole thing, and if you need to restore, overwrite the DB with the copy that was made after restoring. The time the DB is offline for the local copy will be much less than the time it takes rsync to pass through the DB, and will always leave you with a coherent DB backup.
If your connection is slow, and if you are backing up large files, (both of which sound true for you?) be sure to use the keep-partial option.
One of my connections is particularly slow and unreliable. (it's a user desktop over a slow connection) For that one I have made special arrangements to cron once an hour instead of once a day. It attempts the backup, which is often interrupted by the user sleeping/shutting down the machine. So it keeps trying it every hour it's on, until a backup completes successfully. Then it resets the 24 hr counter and won't attempt again for another day. That way I am getting backups as close to every 24 hrs as possible, without more.
Another poster mentioned incrementals, which is not something I need here. In addition to using a version of rsync that does incrementals, you could also use something off-the-shelf/common like retrospect that does incremental but wouldn't normally work for your server, and instead of running that over the internet, run it on the local backup you are rsyncing to. If you need to go back in time a bit still can, but without figuring a way to jimmy in rsync through your network limits.
I work for the Department of Redundancy Department.
Sorry for the self plug, but this just seems silly. Your web host should be backing up your website and offer you restorations. I guess this isn't a standard feature any more. But it is at Suso. We backup your site and databases everyday. And can restore them for you for free.
Hi everyone, I didn't mention in my question that where I'm living (Rio de Janeiro slum) there aren't that many options for internet access. Also, as all my sites are my very much not-for-profit, I'm limited as to how much I can spend on my hosting plan. I've been using Lunarpages for a number of years now, and generally find them very good, although if I stupidly overwrite a file, or want to go back to a previous version of something, I'm out of luck. As I am a lazy (read time-challenged) individual, I tend to use Pmwiki and maintain my sites online, hence my need for regular, physical backups. Anyway, thanks everyone for your help, I still can't help thinking that somebody must be able to make a pile of cash offering such as service to non-techie site owners...
> OK, I keep hearing "use rsync" or other software. What
> about those of us who use shared web hosting, and
> don't get a unix shell, but only a control panel?
As long as you've got scriptability on the client, you should be able to cobble something together. Like, in OS X, you can mount an FTP volume in the finder (Go -> Connect to Server -> ftp ://name:password@ftp.example.com) and then just
(Interestingly, it shows up as user@ftp.example.com in the Finder but the user name isn't shown in /Volumes/.)
AFAIK, pretty much any modern OS (even Windows since 98, AFAIK) can mount FTP servers as volumes. OS X mounts them as R/O, which I always thought was lame, but that's another rant.
> Or who have a shell, but uncaring or incompetent
> admins who won't or can't install rsync?
If you've got shell (ssh) access, you can use rsync. (Not over telnet, natch. If that's all you've got, look at the workaround above.) Rsync works over ssh with nothing more than
Use SSH keys to make life perfect.
Or, google for 'site mirroring tool'. Many have an option to only download newly-changed files.
To get your databases, make a page like
and download that every so often.
For the original poster, who was complaining about downloading many gigs over a slow link, just rsync over and over until its done--if it drops a connection, the next attempt will start at the last good file.
And if you've got a control panel, look for a button labeled 'backup'! My host uses CPanel and there's a magic button.
Final option: how did the data get onto the www server in the first place? Isn't there already a "backup" on your local machine in the form of the original copies of all the files you've uploaded? If you haven't been backing up in the first place, well, yeah, making up for that might be a little painful. (Note: if your site hosts lots of user-uploaded content, ignore any perceived snarkiness. :-) )
Dear Slashdot: next time you want to mess with the site, add a rich-text editor for comments.