Is there a way to run a one way sync?

Support
#1

Hi Everyone, I am just getting started with Duplicacy, so apologies if I am missing something super obvious. I have the web GUI setup on two devices - a desktop and a laptop. I understand the dedupe functionality from two devices, but am wondering if there is a way to periodically sync one of the devices from storage, which has the latest and greatest data? The use case I am trying to solve for is me updating or adding a file on my laptop, which then gets backed up to onedrive. After I get home, I want to run a sync such that the updated data gets pulled down to my desktop (without needing to do a full restore every time).

Thank you!

#2

Duplicacy is a backup software. You need to do sync, not backup. It would be impractical to try to use backup software for sync, it simply wasn’t designed for it, and the resulting solution will be fragile and hard to maintain or trust.

Moreover, if you backup two devices to the same storage they must use different snapshot ID. restoring each devices data to another device may have unintended consequences and hard to track surprises in how conflicts would be handled.

What prevents you from using tools designed for sync to do sync? The OneDrive client itself for example, since you are using OneDrive? Or point to point sync software like Resilio Sync? Syncthing? etc.

And then of course feel free to backup both machines with duplicacy to take advantage of cross-machine deduplication

#3

if you backup two devices to the same storage they must use different snapshot ID.

Is Snapshot ID the same as Backup ID? I am using the same Backup ID across the two devices.

restoring each devices data to another device may have unintended consequences and hard to track surprises in how conflicts would be handled.

Do you mean any conflicts that could be created if that device is backing up as well? Otherwise shouldn’t restore from one device to another should work without any conflicts on unintended consequences? Kind of defeats the purpose of backup if you can’t restore with absolute certainty.

What prevents you from using tools designed for sync to do sync? The OneDrive client itself for example, since you are using OneDrive? Or point to point sync software like Resilio Sync? Syncthing? etc.

So I am using OneDrive as the cloud backup destination vs. syncing tool. The other complication is that all data going to OneDrive is encrypted. So it’s not a drag drop situation. Can look into the other tools, but will they be able to handle decryption from the encrypted chunks on OneDrive to clear files on my drive?

#4

Yes, you are right, in the Web UI it is called Backup ID.

Each machine must use unique (within the same storage target) backup ID:

It is very important. Otherwise you will lose data.

Restore works, but if the existing data at the location you are restoring you onto has also changed – what should the tool do? overwrite with old data or skip? Merge? This is what sync tools are designed to handle. If you are trying to build sync tools using backup tool as transport those are the questions you would need to solve on your own. This is not to suggest to do that – but to illustrate futility of this approach and to suggest to use tools designed for the job.

It’s not mutually exclusive; you can utilize storage there for both purposes.

Great, as it should be, (unless you want web office, etc)

Rclone sync supports client side encryption via crypt remote. Resilio sync supports encrypted folders out of the box as well. Any other sync tool that does not can be made to with Cryptomator or Rclone’s crypt remote.

1 Like
#5

I thought we were supposed to use the same Backup ID so the dedupe functionality can work across devices?

Are you suggesting I create a sync folder on desktop that syncs the encrypted chunks, and then decrypt that folder to get unencrypted files? And I’m assuming I can use Duplicacy to decrypt a local folder. That also means I’m double counting space on my hard drive.

I actually started with Rclone and liked the ease of use and that they have a one-way sync option. Then I read on forums that it’s not great for backups and Duplicacy is better suited. My data doesn’t really change a lot on a day to day basis. I am wondering if Rclone is a better fit for me now. What would I be giving up in using Rclone vs. Duplicacy from backup perspective?

#6

If you want to understand the reasons behind this requirement I highly recommend reading the design paper: Duplicacy paper accepted by IEEE Transactions on Cloud Computing

Either way, it’s always best to read and follow the manual of the tool you are using, especially when data consistency is at a stake.

It is not “vs”. It’s cats vs pizzas.

You see, it’s much better to use a fork to eat steak and use a spoon for ice cream. You can, of course, eat steak with a spoon albeit with a lot of struggle and the same goes for a fork with ice cream. You can also use a single tool like a spork for both to make both experiences equally mediocre. But why?

You have two distinct use cases:

  1. Keep one or more folders on one machine in sync with one or more folders on the other machine
  2. Backup data from both machines.

These two use cases are not related to each other in any way, I’m not sure why are you imposing an arbitrary restriction of accomplishing both of these unrelated tasks with one tool.

For data sync, there are many tools and approaches available: from point-to-point Resilio to rclone mount with crypt via cloud storage at OneDrive.

For backup, you have duplicacy, that supports backup of multiple clients to the same storage thereby saving space at the target due to deduplication.