pCloud, filen.io or rclone

Hello,

how can i using pCloud, filen.io or rclone (booth are included on this) as Backup Storage.

I’m reading about Problems with pCloud and WebDAV like missing chunks and others. For filen.io and rclone i don’t a Way.

A Thread about API Support: Support for pCloud API

Many Thanks!

Many Greetings

Revan335

You can use rclone serve as protocol converter.

I’m not sure I understand what you are asking.

The problems of services in the race to the bottom don’t end with flaky API, if that’s what you are thinking, and your backup is only as good as underlying storage.

Duplicati working with pCloud and Filen include API for 2FA Login.

So can Duplicacy implemented or fix the actually Issues with that too than Duplicati is Open Source.

Maybe can this helping @gchen and the Development to by regulary Updates and Support too. We can using the Power and Help from many Peoples and not only this from one or little Size of Persons.

Hey @Revan335,

I have used pCloud as a Duplicacy backup destination successfully for a reasonably long time (4 years).

I’m not sure I would recommend it unless you already bought into the “lifetime” thing and you might as well use it.

The only thing I’ve found to work reliably is:

  1. Run rclone serve sftp locally with a short dir cache time, --dir-cache-time=5s for instance.
  2. Set up Duplicacy to backup to the local sftp serve.
  3. Only use 1 thread for uploads in Duplicacy, -threads 1
  4. Make sure to check your repo with --chunks as pCloud is unreliable.

Do not use rclone serve webdav even if it’s faster and supports multiple threads reliably. It is a guaranteed way to upload broken blocks.

The first issue with pCloud is that an interrupted upload remains in place, even if rclone correctly specifies in the pCloud API call that any interrupted uploads should be deleted. I have debugged this myself in rclone and pCloud ignores that flag and keeps the half-uploaded file.

Since Duplicacy only uses atomic uploads for sftp, local disk and smb but not for webdav the partially uploaded file remains and the subsequent backup will see it as a chunk that exists. The following check will then flag the chunk as bad.

By atomic upload I mean upload to a temporary file, then rename on destination when the upload is complete. (@gchen if you’d be willing to implement a “pedantic” mode that always uses temporary files for upload, like Kopia, that’d be super, but I’m not expecting we work around bad cloud providers :slight_smile: ).

The second issue with pCloud is that sometimes when you go and download a file, it’s just not available. Rclone will return a “server not available” or “file gone”, sometimes for days at end. The data has so far always been available in the web UI (pcloud.com) but this is obviously not tier 1 behavior.

Now saying all this about pCloud, I have still successfully backed up countless machines and restored and recovered data successfully many many times. Duplicacy’s robust server-side storage format makes it trivial to recover from most issues, something that can not be said for all backup software.