But I decided to only use Borg’s history settings. Syncthing also has some history features, like keeping versions, or sending deleted files to an intermediate trash folder. Besides a 50Gb disk is way more than I need and only costs a couple of dollars per month.īackups serve 2 purposes, extra redundancy and lightweight historical snapshots. The fact is that Syncthing does a lot of scanning, which just doesn’t play well with the unorthodox nature of a HTTP REST-based API file storage system mounted onto a local disk.

However, eventually I gave up on it as it caused some frustrating performance issues in Syncthing. I wanted to not have to care about resizing my own disks, with a bucket you essentially have infinite pay-only-for-what-you-use storage. I originally spent about a month experimenting with mounting my remote VM Syncthing folder onto a Google Bucket (just like AWS S3) using gcsfuse.

When Syncthing says that a folder or device is in a certain state such as “Up To Date”, this just means locally, as in Syncthing is aware of the local state and is ready to transmit the metadata of that state elsewhere. The Syncthing design consists of just 2 concepts Folders and Devices. There are of course lots of config settings so you can do things like mark a folder as Send (or Recieve) only. Every install can sync to every other install and vice versa. Syncthing is peer to peer, so there is no client/server model.

There is also clipboard sync so that I can copy/paste between my local tmux environment and my remote environment. The files are mostly backups of things like photos, wifi passwords, 2FA config, SSH keys, laptop config and ZSH history. I use Syncthing to synchronise certain files between my phone, laptop and remote VM. Here are some brief notes about installing Syncthing and Borg.

Syncthing aws s3