This topic was started in Github Issues
I will explain it again here.
The major problem, why this topic exists is because of the fact, that Crashplan will crash in october. The only alternative (at least, what I could find) is BuddyBackup
BuddyBackup would replace Crashplan completly, but it has the following problems:
Lately it’s central server has been down, without any notice.
There is no response from the email on their site neither from their parent company Databarracks LTD, making the downtime and the future of the program more concerning, that it will in one moment just not work (also rendering all backups useless).
Last update: 2014, tutorial with WindowsXP screenshots.
I will also mention, that there are backup solutions, but they require monthly or yearly subscription for the service or for cloud storage. That will move us to making your own offsite machine or friend’s machine. The closest to that is P2P: Resilo & SyncThing. SyncThing is chosen, because the open-source (if it closes down, can be made operable by individuals) and the freedom to donate instead of paying monthly (tough paying developers monthly would help loads for those who can afford it). SyncThing is the best variant, what has most built.
That brings us to SyncThing and it’s currently missing features:
Encryption: (since how it works, it goes along with compression): If not backing up to really trusted other side (offsite dedicated server, husband’s desktop), then it’s essential, that the other side doesn’t have access to your files. Password should be printed out or chosen by user, in case of failure or destruction of the pc what is being backed up, the files can be decrypted.
Compression: When backing up, it’s not important (sometimes it definitely not be), that the other side can access the files quickly, so saving space is a priority over accessibility. The files should be compressed and encrypted by chunks, since any changes or additions would need the whole thing (on most ways) to be recompressed and reencrypted.
Sync to external drive, when plugged in: This is useful for onsite or offsite backup for single users (not everyone is gonna buy a cheap pc and drives and leave it running offsite).
Since the users of Crashplan are already worried, they are seeking a solution. In october, when Crashplan shuts down, users will have to find a new solution.
BackThing or BackupThing (I prefer BackupThing, but @uok in github issue liked BackThing as a name more) should be functional to replace Crashplan. Also with that would come many devleopers and donations. I will definitely donate to this and I also want to learn Go and afterwards help to develop this.
What do you all think? Anything to add or remove from the current plan?
I’m also a long time CrashPlan Home user and looking for a reliable alternative.
I’d rather pay/donate for a developer and open source software than add a subscription for yet another service and trust an (US?) company with my data (like Dropbox that only encrypts data in transfer but not the data at rest)
multiple platforms (Linux, Windows, etc.)
easy to use
backup to remote device (own or friend)
backup to remote storage (Amazon S3, etc.)
no (but possible with own server)
no (but often requested feature - see bountysource)
versioning with restore
no (Syncthing only keeps versions of local changes)
Actually you can back up to remote storage (AZ3, GCS, Azure and more), when you buy some software (CloudXtender, expandrive)
I totally forgot, that you need a feature to assign a share to any local folder, not just a general location in settings. Otherwise you can only use SyncTrayzor to backup to only one drive. Sure you can merge the drives, but that requires some software, what again you need to buy or merge the drives with in the process deleting them. I don’t right now know, if Linux (mainly Ubuntu) has a feature to merge the drives, without erasing them, because you can always make your offsite run on a linux distro (tough when you are syncing to a partner (friend), then ‘install linux’ wouldn’t make sense)
I was going to propose: Use Borg to do the backup work. You get compression. If you choose to, the backup may be encrypted. It’ll definitively be compressed. It’ll be chunked, so only changed blocks (=fewer changed files) will be replaced in the backup store.
Then you may sync the backup itself to an off-site computer using Syncthing. On the off-site, you may even locally backup your backup…
Personally, I wouldn’t do this. Remember, Syncthing is a continuous synchronization program. If you delete (some part of) the local copy of the backup, then (some part of) the remote copy will be immediately deleted too. And don’t rely on the versioning functionality of Syncthing to work around that. The versioning functionality of Syncthing is not meant as a backup system (similarly, the Trash Bin of your OS is not considered as a backup system).
Backuping a backup is a sign that your initial backup is not a backup…
Borg is nice and seems mature, but can only push to a remote location over ssh if I remember correctly. If you need to push your backup to some cloud, projects like duplicati, duplicity or restic seem to propose that. (Note that I don’t make any statement about their respective maturity or quality, as I only have tested duplicity and duplicati).
Anyway, the TL;DR is: Syncthing is awesome at doing what it is meant to be (continuously syncing files), but don’t use Syncthing as a backup software. (and personally I add: don’t even use it as part of your backup strategy).
Backups are meant to be robust and hard to delete/corrupt/overwrite.
@zertrin, @Guz thanks for commenting, but the topic creator suggested a NEW software that is based on parts of Syncthing to create a backup alternative that is easy to use like CrashPlan Home. The idea behind this thread is to attract CrashPlan users who can also contriubute as programmers to this new software idea (and Syncthing)
As a Crashplan user (both as a private user and at work) I would be interested in this. As uok said - it was clear from the proposal that this is a discussion about new software, in the spirit of Syncthing but not Syncthing. Features more or less similar to Crashplan would be my preference - (un)limited versions, store everywhere (network, local disk), define backup sets, start-and-forget.
(As a side note - not only will Crashplan stop supporting private users, (hm, I wonder what happens with the four-year license I bought…) it’s getting increasingly more difficult to use at work as well. Our enterprise version (which just means that the ‘cloud’ is to be provided by ourself) is getting worse with each iteration, and now Crashplan tell us that “it’s not intended to be used for servers, only desktops”. So we’re moving away to alternatives there as well.)
ok boys and girls, I really wanna make this happen - at least for Syncthing first!
As I see it the big missing feature is the encryption so that data can be stored on (untrusted) servers. @calmh, @AudriusButkevicius can we hire someone to implement this? What are your thoughts?
Certainly I expect it’s possible to do so. I’m not really sure what you’re asking, I think. Are you asking if the project has the funding to hire someone at commercial rates to implement this? We do not.
Remember that backup programs are hard to write. Your job is to delete data, but only the right data. If you screw up, most people won’t notice until it’s too late. People aren’t going to use your software until you can safely say “this is mature enough to use in production, and to trust your data to”. That makes testing hard. You can’t say “use with caution: keep backups”, because that defeats the point.
Also remember that good software takes time to write. I reckon Syncthing’s had at least 2 man-years of effort. Borg is similar. Make sure it’s what you want to be doing for a while . If you want to pay someone else to do it, you’ll need investment and a business plan. It’s going to be cheaper to subscribe to someone else’s backup plan. It’s also going to be a while (years) before something’s ready: will the Crashplan users have found something new by then?
If that’s your only complaint, it’s going to be an awful lot easier to modify borg to do what you want, than to write an entirely new piece of software.
Don’t be a dick. Regardless of everything else it’s valid to discuss what could or should be done separate from who will be doing it. Not every feature request needs to be backed up by the ability to solve it themselves. A lot of the backup specific things mentioned fall into “should not be done” imho. Encryption is not one of those things.