Issues while accessing repository through sftp

Hi all!
I am using restic 0.17.3 compiled with go1.23.3 on linux/arm64 to transfer my backups to a remote repository (Hetzner).
I am using sftp as transfer protocol.
This morning I wanted to run a restore test, but I was unable to browse my snapshosts:


root@server:~# restic -r sftp://xxxxxx@xxxxxx.your-storagebox.de/ mount /mnt/restore
xxxxxx@xxxxx.your-storagebox.de's password: 
enter password for repository: 
repository dcb067c6 opened (repository version 2) successfully, password is correct
Now serving the repository at /mnt/restore
Use another terminal or tool to browse the contents of this folder.
When finished, quit with Ctrl-c here or umount the mountpoint.

on another shell:


root@server:~# ls /mnt/restore
ls: reading directory '/mnt/restore': Software caused connection abort

I have a couple of questions:

  • it would be better to use a different protocol (REST, S3,…) instead of sftp, in order to avoid timeout?
  • my snapshot contains different data: some big file (Veeam backup), some mysql dumps, some file system dump etc… it would be better to store each “category” in different snapshot, or is it ok to put every data in the same repository? I thought that by accessing a snapshot that contains LESS data, I would have fewer problems

Thank you very much!

Directory metadata is cached by restic, so the protocol doesn’t really matter.

Performance is primarily determined by the overall size of the repository. The size of individual snapshots barely matters. (unless you have TB sized files lying around)

Did restic continue running after that? Does the server have enough free memory?
I’ve never seen this particular error code. Googling for it looks like a windows error most commonly seen together with putty or samba??? Neither makes any sense based on the setup you’ve described so far. So consider me completely confused as to what’s going on.

Hi Michael, sorry for my delay. My snapshot is about 3 TB, and it contains 9500 files. I am accessing my remote repository through a Linux host. After getting that error (“Software caused connection abort”) I tried again, but… same result.

Now I am trying to “split” my backup into multiple snapshots, so that every snapshot contains fewer data.

I will update with the results :slight_smile:
Thank you for your answer!

That very likely won’t make a difference unless all the files are part of the same folder.

What about those questions?

Hi Michael, what’s about “the same folder”? Do you mean “same repository?”
Now I am trying to upload a new set of data (as I wrote: not a single repository, but multiple repositories). After that, I will try mounting again and I will update you here :slight_smile:

Thankyou