Ask Slashdot: Syncing Files With Remote Server While On the Road? 239
An anonymous reader writes "Here's a scenario: you are on a vacation trip for a couple of weeks — on the road. Lots of pictures — 2-300 per day. Maybe some text files with short notes etc. You have a camera with Eye-Fi, a PC, and a phone with WiFi and 3G. Files ends up on the PC (mobile storage), phone provides Internet connectivity. Now, if you wanted to upload all files pretty much as you go — given spotty access to Internet over G3 and WiFi — what would be the best way to do that automatically; set-it-and-forget-it style? I would like them to end up on my own server. rsync script? ownCloud? Some BitTorrent setup? Other? I'm thinking of interrupted file transfers due to no network, re-starts etc. And I would not want to lose any files; including scenarios where files gets deleted locally — that should not result in files getting automatically deleted on the server as well. Sure; I could perhaps use something like Dropbox but that would take the fun out of it."
So having us piece something together for you (Score:5, Informative)
Leaves the fun in it? Either take the suggestions that work, ie Dropbox, or figure it out yourself.
git-annex - "The Nomad" use case (Score:5, Informative)
http://git-annex.branchable.com/
use case: The Nomad
Alice is always on the move, often with her trusty netbook and a small handheld terabyte USB drive, or a smaller USB keydrive. She has a server out there on the net. She stores data, encrypted in the Cloud.
All these things can have different files on them, but Alice no longer has to deal with the tedious process of keeping them manually in sync, or remembering where she put a file. git-annex manages all these data sources as if they were git remotes.
When she has 1 bar on her cell, Alice queues up interesting files on her server for later. At a coffee shop, she has git-annex download them to her USB drive. High in the sky or in a remote cabin, she catches up on podcasts, videos, and games, first letting git-annex copy them from her USB drive to the netbook (this saves battery power).
When she's done, she tells git-annex which to keep and which to remove. They're all removed from her netbook to save space, and Alice knows that next time she syncs up to the net, her changes will be synced back to her server.
Fish-Sync (Score:3, Informative)
There is an open source tool for just this kind of situation. It is called Fish-Sync and can be found here: http://fishsync.sourceforge.net/
It basically works like Dropbox or other sync services, but it syncs files between computers you have access to, rather than a third-party server. Fish-Sync is basically a combination of rsync + OpenSSH + Dropbox's LAN sync with an optional pretty graphical interface. (It also works on headless serves if you don't want the GUI.
Re:Take fewer pictures (Score:4, Informative)
Oh, any unreasonable photographer will do that in a nice place.
That said, uploading to the Internet is kinda dumb. Just take a pair of cheap hard drives, download it to both, keep a copy on the laptop. If you're going through customs, give one to someone else in the unlikely event you're mistaken for a terrorist. You can also mail a drive back.
Forget the Internet for a while. You'll be healthier and happier.
Unison (Score:5, Informative)
I am a recent convert to Unison. I discovered this because I was trying out Google Drive, which I found worked well, but I don't want to keep more of my data in someone else's cloud if I have to. So far I have been using the Mac OS UI. With a solution with Unison you will need your own server with ssh access to the Internet. The downside is that you have to worry about backups or uptime, though you don't have to worry about some government taking the service offline permanently (or while they spends years trying to establish possible guilt).
The advantages with solutions like Drop Box or Google Drive is that you not have to worry about the server side. Depending on the amount of data you want to store you will have to choose between the basic paid access or the paid access, which gives you more storage.
BitTorrent is probably the worst solution here, since it only works well when the data is massively distributed. If you only ever have one peer, then you are better off with one of the solutions mentioned above.
Re:Let me be the first one to say (Score:4, Informative)
Why use cloud service like dropbox when you can do what real men do and build your own using unison -
http://www.cis.upenn.edu/~bcpierce/unison/ [upenn.edu]
Re:git-annex - "The Nomad" use case (Score:5, Informative)
I've been using git-annex for this kind of synchronization for a while and have found it works quite well. It supports interrupted file transfers (via rsync), automatically keeps hashes of all your content, and lets you work offline effecively. git-annex is particularly good at synchronizing file renames between two machines: it actually tracks the renames so you won't have to re-upload anything like you would with rsync. It also supports more than one machine, if you want to make backups.
I would definately recommend trying it.
Re:So having us piece something together for you (Score:5, Informative)
Dropbox probably isn't going to work.
Yes it will. Perfectly, actually, in my experience.
And it won't delete photos when you delete them from the camera upload directory.
Check it out here: https://www.dropbox.com/help/288 [dropbox.com]
Further, it will chew on each image file till it gets a successful upload.
Try CrashPlan (Score:5, Informative)
I've been running CrashPlan as an online backup solution for my Linux system to back up all of my photos. It also has a feature allowing you to back up to another PC over the Internet. It's easy to set up so you can back up to your home PC and it's free (unless you buy the cloud backup service). See http://www.crashplan.com/ [crashplan.com]
-Aaron
Rsync; better yet: datamover (Score:4, Informative)
I think rsync pretty much provides all you need in one tiny command-line to get data from A to B.
But if you want to increase your resilience against failing network connectivity, and make sure you don't delete anything that hasn't been properly copied to your server, I suggest you take a look at datamover: http://www.cisd.ethz.ch/software/Data_Mover [cisd.ethz.ch]
Essentially, it's a daemon written in Java that monitors an outgoing directory. Everythings that is written in there gets safely copied over to a central storage drive. Behind the scenes, they use rsync to do the copying, but it's wrapped in tons of features that improve the reliability of the moving process, like a quiet period before a file gets moved (good for applications that write their output incrementally and sporadically into files), multiple retries on network time-outs, high-water marks, data transformation (e.g. compression) during the move process, etc. It also is very anal about sending you emails for anything that could possibly be a data integrity problem.
We rely on it to store the raw data from scientific experiments. With the proper configuration, your holiday pictures should be just fine.
Re:So having us piece something together for you (Score:5, Informative)
rsync makes no sense to me, because of the requirement that file deletions on the mobile device should not be repeated remotely, which is exactly what rsync normally does.
It doesn't delete files unless you specify --delete
Re:Take fewer pictures (Score:5, Informative)
Not true. This from an interview with Adams' son:
Ansel Adams frequently made duplicate photographs of his images when taking them. One thing that I tell people constantly is that it is always a good idea to take more than one shot of an image if you can in the camera. According to Michael, Ansel frequently took multiple exposures of the same shots. Many of his negatives are duplicate images of which he'd select the best image to use for printing.
10 Interesting Things I Learned About Ansel Adams [flickr.com]
Re:Fun? (Score:5, Informative)
You say you don't want to take all the fun out of it, but you're trying to foist this idiocy off on a public forum? Save the fun for yourself, and make a blog post about your solution.
Allow me....
Hello from my vacation in Indonesia. There is no automated solution. What you're ignoring is that networks in the developing world are not only patchy, they're flaky too. So, whatever worked for you yesterday might not work tomorrow or even in half an hour's time. Counting on 3G is a bad idea, because of its unreliability, but also because of its cost. Use wi-fi wherever you can. Most hotels these days provide it free of charge. Use rsync (with the zip option if you like), and keep it simple. My update script looks like this:
rsync -av ${SRC}/* ${USERNAME}@${DEST}:${PATH}
Yep, just a single folder in which I dump everything of value and a corresponding folder on my home machine. I just pop open a command line whenever ity's convenient (and possible) and run it. It doesn't always complete in time (the one I'm running as we speak won't be finished before I leave to go scuba diving), but I can always complete the sync later in the day.
Also, bring one or more external disks. Use them for quick and dirty backup while you're on the move. It only takes one rain storm (or fall in a river) to be glad you did. And don't count on buying new SD cards when you're on the road. Most of the ones for sale in the developing world are convincing knock-offs that last about two weeks. That's my experience anyway.
Anyway, simplicity is the single most important step for you when you're backing up data in the developing world. You can't rely on any other factor, so you should at least be able to rely on your own scripts. Which leads to my maxim: "In the absence of robustness, choose simplicity"
Re:Fun? (Score:5, Informative)