This comparison site (http://www.cloudwards.net/comparison/) tells me that http://www.adrive.com is the only free solution with +25GB and support for Linux. It has no reviews yet though.

Maybe I have to start to pay for storage.

This list (http://www.cloudwards.net/comparison/) tells me that the only free option providing more than 25GB, for linux users is http://www.adrive.com/.

There are no ratings of this service yet. Maybe I will start to have to pay for storage?

Faster, in the sense of recovery. Backing up is about the same speed as using the cloud (a cloud service, such as Carbonite), but restoring a disk crash is hugely faster — because it doesn’t depend on my data dripping back into my machine via the internet.

“Just add Linux.. and go!”

Did I use the internet to restore his system? Are you kidding? Wanna do the same with your 5 TB held by Carbonite? Are you kidding? Here’s where my scheme has also paid for itself…

Since all our servers are built with RAID-1 drive arrays, I simply split his mirrors and drove over to his house with his mirror disks. Then I removed the mirrors from his corrupted server and swapped in the mirrors from his backup unit. After bringing up his server and restoring its identity (machine name & IP address), I re-added his useless corrupted disks to the “transplanted” RAID array.

The server was immediately usable, and started healing its array — which took hours to complete. (Transferring the data over the network would have taken WEEKS.) When I got home (and his machine was 100% again — so that his data was safe), I did the same with his backup machine; soon it, too, was back to 100%. Both machines were usable the entire time, except when I was physically swapping disks.

No proprietary backup software needed, no backup service to pay for, no slow restores over the internet, very little downtime, I’m in full control, it’s fully secure, and we can log into our backup machines to retrieve individual files to restore. (I’ve had to do that a few times myself.)

I plan to change the nightly ‘rsync’ mirrors to a COW scheme — Copy On Write — using something like ‘rdiff-backup’. That will create a system like Apple’s Time Machine, that will keep older versions of files, too, not just the latest (which are, at most, 24 hours old). Currently, if you don’t recover your lost/corrupt file within a day, it’s gone. And of course, it may take more than a day to realize you need an older version of a file…

I’m doing just what you described: My system at home (actually my file server at home; I have multiple PCs & laptops) mirrors itself every night over an encrypted link to the machine in my parents’ house. I use an ‘rsync’ cron job for that, which runs at 1am.

And, since I’ve also converted my parents to Linux (when Ubuntu 8.04 debuted), I have a cheap, old PC set up in my house that does that same thing to mirror their desktop PC. Their laptop backs up to their desktop prior to the desktop doing so to my house. (Actually, the machine I set up for them receives mirror backups from my in-laws’ machine, too.)

And, since I converted my best friend to Linux some years ago, and he’s running a business out of his house (pro photographer), I built him a Linux server with about 6 TB of RAID-1 storage. It mirrors to a similar PC in my house every night, too.

We all have fast cable modems, with 50+ Mbps down, and 5-10 Mbps up, so unless my buddy shot GB of images that day, everything “clears” by morning. If not, it keeps transferring throughout the day. Since the bottleneck is the upload rate, no one’s modem is choked; receiving backup data takes ~10% per machine, and backups that extend into the day are uncommon.

(On the rare occasion that it takes > 24 hours to complete a backup transfer, the cron job kills any existing session first, then restarts a new one; ‘rsync -P’ simply continues where it left off…)

For security, I use a nifty open-source stealth port knocker called “FWKNOP” (FireWall KNock OPerator) that requires an encrypted packet to silently open an obscure port on the other end just long enough to connect an SSH tunnel. Rsync, of course, tunnels through using SSH…

http://www.cipherdyne.org/fwknop/docs/fwknop-tutorial.html

I had been thinking of setting up with someone else, like my father in law three towns over, to share a backup system, encrypted drives at both locations mirroring with a simply intertoobule linkup.

Anyone who’s into “instant gratification” and loves today’s fast machines will be hit with the dissonance of “relax, and go get a beer while you’re waiting” if they try to use the cloud as a “convenient universal hard drive” for daily work. Mostly so for big files…

Backing up is probably fine — because you want that done at night, while you sleep, anyway…

But in my case, it’s over an encrypted internet link to a machine that *I* built and control, not to a public cloud service. That meteor would need to make a crater more than 100 miles in diameter to hope for a chance of knocking out my house and my backup site.

But I do things that way because I can…

Second use, also important, for a subset of what I’ve got, to have a few directories mirrored on two or three machines that I move between.

Third use: To share a directory with a bunch of files, or a file, with someone.

I have nothing in the cloud I don’t have on my main hard drive and a second backup hard drive. Both of which would be destroyed if my house is hit with a meteor. Thus, the cloud. Hopefully it will not be too big of a meteor.