John Eberly

There has been numerous posts for online backup solutions (for good reason), but I haven’t found the best solution for me yet.What I want: cheap reliable off-site backups for my home Linux (Ubuntu Server Edition) file server.

O.k. I am a classic digital packrat. My actual garage is very clean and organized, but my digital garage is full of every byte I have saved since 1998, old Job emails, photos, music, websites, some videos, etc. I never, ever want to lose this stuff, even if my house burns down because my laptop battery explodes. So I need to backup my stuff, simple huh? Well, I already store all of our files on a Linux file server via Samba. This machine is backed up nightly to my Linux Desktop machine with cron and rsync, but this doesn’t solve the fire/flood issue. I used to copy data to DVD and store off-site, but burning DVDs is a pain and my data is out growing DVDs. I currently have approx. 10GB of data, but expect it to grow, especially since I want to store our digital camcorder footage as well. So I began the quest for remote storage. While these are not all of the options I considered (I looked/tried many) they came out at the top of the heap for me.
My dream remote storage.

1. Cheap, of course ($10 or less per month)
2. Secure, some of my files have info I don’t want in the wrong hands.
3. Relatively fast, but I am sure my upload speed (768 k) from my ISP, Comcast will be the bottleneck.
4. Some sort of Linux server of with shell access (so I can automate with cron and rsync)
5. Not have to set up another full fledged energy hog of a server
6. Bonus, unlimited scalability.

How to store files offsite? Let go searchin’…….

Option 1: Amazon S3
This was actually my first choice, but I haven’t been able to find a good tool to use the service. I tried JungleDisk (on linux and windows), but JD had so many crashes I decided to never use it again, but I am still interested in S3. I have heard people express concerns of whether Amazon S3 storage will be around forever and I cannot answer that. While I do think Amazon S3 will be successful, I do trust Amazon over most to provide adequate time to move off of their service if they ever decide to discontinue it.

Pros:

• From Amazon, a big name company with a very good reputation
• Relatively cheap at $0.15/GB stored and $0.20/GB transferred
• Infinetly scalable and stored by Amazon in geographically distributed data centers (huge plus)
• Reliability… I assume this would be the most reliable of all my options

Cons:

• No simple way to use rsync to upload data

Summary: I just wish there was a reliable tool to get my data uploaded like rsync/ssh. s3sync looks promising but I haven’t tried it yet.

Option 2: Dreamhost.com
After researching S3 for awhile, I thought maybe I should just get another hosting account. Dreamhost had everything I wanted, shell access, plenty of storage (200GB) and only $40/year after googling for coupon code. Sounds to good to be true. 1and1 had similar offerings, but I have heard to many horror stories to try 1and1.

Pros:

• Cheap
• Relatively good reputation
• Shell access
• Ability to host websites there as well

Cons:

• Security concerns

Summary: Well they obviously oversell, they even admit it. But that didn’t bother me, if I could use the 10-50GB I would need. So I signed up and started to rsync some of my data, but then I started to worry a little. Anybody with $40 could sign up for an account with shell access. What would happen if someone gained access to my files? After a simple ls of the home directory, the first site I looked at that was hosted on my server was a blog from an ethical hacker, I decided to jump ship entirely. Not that I think Dreamhost doesn’t secure their servers, I am sure they do, but it just didn’t feel right to leave my somewhat sensitive data there. I would trust it for my photos and videos, but not my documents. So now what.

Option 3: Modified Linksys WRTSL54GS router with openwrt or dd-wrt firmware and external USB drive.
I thought about taking a Linksys WRTSL54GS and putting the much improved firmware options available from openwrt or dd-wrt and slapping a USB drive on top and then putting it at my mom’s house. This would cost around $200 for about 250 GB and a router. This one appeals to the DIY part of me. I would set it all up and make initial data sync locally, then mail it to her to save a lot of time with the initial upload.

Pros:

• No monthly fees
• I assume this setup would consume less energy than a traditional PC, but I have not calculated anything yet.
• All hardware is under my control
• fun to set up

Cons:

• One more piece of hardware to buy and maintain, secure etc.
• Limited to upload/download speed of my mom’s DSL
• consumes her electricity

Option 4: Carbonite.com
Pros:

• Easy install on Windows, “unlimited” for $5/month. Thought I could use with Samba share, I was wrong, shares not supported.

Cons:

• Doesn’t work with Network shares

Summary: Since it doesn’t work with network shares, this is not going to work as a solution for me, but would be a very good solution for the average Windows user with photos, music etc in his My Document folder.

Option 5: Use my extra space on my webhost ~ 10GB
Although this is a tempting option, I don’t want to mix my hosting with my backups. If I accidentally filled up my webhost quota, I could risk downtime. I ruled this option out right away, but I thought I should mention it.
Conclusion

Here are my cost guesstimates for two scenarios….

Assuming 2 years and 20GB of storage

	Storage Costs	Transfer Costs	Hardware Costs	Total for 2 years
Amazon S3	$3/mo	~$0.50/mo	none	$84
Dreamhost	$8/mo	none	none	$192
Linksys/USB drive	none	none	$200	$200
Carbonite	$5/mo	none	none	$120

Assuming 2 years and 100GB of storage

	Storage Costs	Transfer Costs	Hardware Costs	Total for 2 years
Amazon S3	$15/mo	~$2/mo	none	$408
Dreamhost	$8/mo	none	none	$192
Linksys/USB drive	none	none	$200	$200
Carbonite	$5/mo	none	none	$120

So what am I going to use? Good question. First of all, Carbonite will not work for me at all. No way to backup from Linux or even Samba share. The Linksys/USB solution would be kind of cool, but is one more device for me to have to maintain and it will be offsite. Plus, it will be limited to my Mom’s DSL upload speed, if I need to restore. It would be faster, literally, to overnite the drive to me, than to download 20-100GB back.

I am left with Dreamhost or Amazon S3. While Dreamhost seems like a good fit for me, I am very nervous at the thought of letting my data sit on a server that 500 people have shell access to plus not very scalable. So I am left with Amazon S3. I only have a couple of reservations about this. One, privacy of my data, I am trusting that completely to Amazon, for better or worse. Secondly, I am worried I might start to backup much more data which would push the monthly price over my $10 monthly limit.

My next post will be on the success or failure of my Amazon S3 via s3sync option. Stayed tuned…..

Leave a comment if you have other ideas on how best to solve the growing challenge of offsite storage.

While I have been researching this for awhile now, I was inspired to write my findings down after reading a good post on this topic from Jeremy Zawodny and realizing many people are interested in this topic.