This message is for all the developers of restic, I just wanted to reach out and thank you. I’m sure there are moments of frustration dealing with the inevitable issues that pop up, but it is a brilliant piece of software, well thought out and incredibly useful!
Also, I wanted to thank everyone for their continued consideration of edge use cases. So often the answer is “you’re holding it wrong” It’s incredibly refreshing to see the project always open to consider new needs and approaches.
For example, I was stunned at the massive improvement using the --ignore-inode flag for backing up a fuse mounted file system. We use the encrypted file service Tresorit which allows you to mount your directories and files via fuse mount. Wanting to back these files up to Backblaze I naturally turned to restic as I’ve had so much luck with it. There are something like 600,000 files totalling 450 GB on this mount. Using restic without the flag each backup was taking nearly three days. I did some research and found the --ignore-inode flag and it’s now taking under 20 minutes.
So we have an encrypted fuse mount of nearly a half TB and 600k files on a VPS that only takes up a few GB of actual hard drive space, that’s able to be incrementally backed up to a encrypted remote B2 repo in under 20 minutes. That’s incredible.
To top it off when I found that flag I was about to remove the apt version of the restic program and manually add the binary, as that version didn’t have it yet (Debian buster uses 0.9.4). Instead you guys made a simple self update function that worked instantly. Great work.
I just wanted to +1 this.
I have been experimenting with a few stacks over the years to find the best backup solution for my (personal) needs.
And I just love Restic, the de duplication works great!.
here are some stats from my backup (to B2 - love that Restic supports it, in fact this led me to Restic to begin with)
scan finished in 405.026s: 192265 files, 307.951 GiB
Files: 192265 new, 0 changed, 0 unmodified
Dirs: 30880 new, 0 changed, 0 unmodified
Data Blobs: 166956 new
Tree Blobs: 26007 new
Added to the repo: 130.476 GiB
processed 192265 files, 307.951 GiB in 15:17:06
this is an amazing saving rate!
I’ve also played a bit with restic mount and I was really happy to find out
that I can mount a restic repo, find the snapshot with an encfs encrypted directory and then mount it directly from restic mount
so I could access an encfs (on the fly) decrypted file without downloading the entire encfs dir!!
this is just awesome, thank you so much.
I would to see the compression implemented on top of this, if it will end up saving even more space (I guess the compression can be done on the blobs after hashing them?)
Yes, developers need hi fives. Thanks for restic. I was fooling around with borg backup before but it was so complicated to setup. If backups are complicated, people won’t set them up! (I never did!)
I now have a local restic backup on a hard drive, and an offsite rest-server backup at another location. So much easier to use than borg! I even wrote a tutorial for setting up Restic with Yunohost because I love it so much.
I discovered restic a few months ago and started using it to backup my laptop files to an S3 bucket. Today I had to migrate to another laptop so I restored all the files from my home directory to a fresh Ubuntu installation. Worked like a charm, and the speed was amazing!
Also the ability to restic mount a repo to the local filesystem and browse files is a killer feature. I plan to transition my other machine backups from Duplicati to restic.
I am also learning restic to leave Duplicati since it has given me some problems with the database.
What are you talking about with “browse files”? Some GUI browser por Restic like “Restic browser”?
Browsing your backup as a regular file system is also very easy. First, create a mount point such as /mnt/restic and then use the following command to serve the repository with FUSE:
…
With rclone it is possible and I have used it without problem to mount a Google Drive folder as a virtual drive. Hence my doubt about restic and WinFsp
https://winfsp.dev/
"WinFsp is system software that provides runtime and development support for custom file systems on Windows computers. In this sense it is similar to FUSE (Filesystem in Userspace), which provides the same functionality on UNIX-like computers.
Typically any information or storage may be organized and presented as a file system via WinFsp, with the benefit being that the information can be accessed via the standand Windows file API’s by any Windows application."
If you have issue not related to original thread subject open new one.
restic and rclone are two different programs
as @KamikazeePL said restic mount does not exist on Windows - you could check this in documentation
It could be possible to implement it. Either you do this and post PR (beauty of open source) - many people will be happy. Or you request new feature and wait until somebody does it (maybe tomorrow maybe never). Check first if it is not already requested on github,
Thank you for restic !
Use daily for the last 3 years with different backends ( azure,aws s3,oracle oci,bb2,pCloud,mega,google,dropbox,sftp etc etc)
Size of repos 1Tb - 10Tb. No problems at all ( well - very minor and with workarounds ), restic is getting more mature and it is such a pleasure to use it.
Thank you again.
Janek
Hi there, I am new to this forum and as the very first thing I want to thank you for the fantastic work you have done with this program.
After lot of reading I decided to give it a try … normally I was quite happy with my borg setup.
What I learned so far is that borg and restic are quite similar and it’s merely a matter of personal taste what to use, when you don’t depend on a unique feature like AWS S3 support or so. In my environment (all linux servers, mostly debian) both are working equally good. The one thing making me consider to switch over to restic is that for my personal use case the locking of the repository when mounting it somewhere on the target host is more flexible. So I was very happy that I can add snapshots on the fly without the need to unmount the repository.
I am just using it a few days now, performance seems to be a bit better but nor significantly.
Again, I can only say thank you for the good work and I am happy to follow the progress here.
Restic is neater than Borg. The repository format is specified , so everyone can fork and continue the project. Borg is written in Python and has some dependencies such as low level C libraries. You can backup the restic static binary with repository and recover the archived data long in the future. The command line options are also a bit messy.
The encryption in Restic comes from a better trusted library.
Borg has no integration with cloud. You have to keep a local copy, and sync that (unless there is access over SSH).
Borg is single treaded due to Python GIL, though it seems fast considering that limitation. It may be overall better or worse than Restic in terms of speed and RAM usage, depending on the data.
Overall, Restic is newer and better in my opinion.
I didn’t want to start a discussion here about pros and cons of borg compared to restic. In my environment, they are pretty much the same, I even didn’t use the borg encryption, since all my storage is on encrypted LVM volumes anyway. Since I’m new to restic and in the tutorials there were no easy way to create volume without encryption (there surely is :-)) I use it now, but it’s no key feature for me.
Borg with ssh and restic with sftp feels the same, I know there is a difference, yet I did not install a rest-server, which would be in my understanding more comparable.
For now I’m happy with the fact, that I do not need to care, if something is installed on the receiving site. (arch<->debian … very different versions).
I do not see so much difference in speed (mainly limited by network for both) and RAM usage.
I think i should mention, that the repository format is also very well specified in borg … that is a must, I would not trust any software hiding the specification of data storage.
Since I switched over I had no issues with restic so far, it gives me a good feeling, so I will continue to use it for new snapshots, Since I have a huge borg repository with historic data and I have to admit I’m too lazy to think about a migration strategy I will continue to use both of them.
It’s incredibly refreshing to see the project always open to consider new needs and approaches.
