Downloads von Backups werden beim Klicken auf den Link nicht gestartet

hi,

since today I can not download my created backups via the e-mail link anymore.

  • tried different firefox based browsers, but this shouldn’t matter as it worked always before
  • tried to download also older backups from the backup list in discourse
  • updated docker_manager and discourse itself to the newest version today (hosting the discourse docker instance on a vps, never hat any issues)
  • restarted the vps two times in total by now
  • did one ./launcher rebuild app rebuild

what happens: I click the download button in discourse and discourse sends me the download link via mail. if I click it, it loads quick but the download is not being initiated. if I reload the new tab, it tells “Sorry, the backup download link has already been used or has expired." Even copying the link and opening it in another browser doesn’t work

Discourse Version: v3.6.0.beta1 +37
Docker_manager Version: 78a2036b

What do you get if you try wget -v YOUR-BACKUP_LINK instead of downloading via browser?

Resolving the hostname <MYDOMAIN> (<MYDOMAIN>) <IP address>
Connecting to <MYDOMAIN> (<MYDOMAIN>)|<IP address>|:443 … connected.
HTTP request sent, waiting for response … 404 Not Found
2025-10-01 10:22:33 ERROR 404: Not Found.

But all backups seems to be available here: /var/discourse/shared/standalone/backups/default , the names and the sizes of the .tar.gz files are correct.

The Links look like this (I replaced two characters with two * in the token):

https://<MYDOMAIN>/admin/backups/<DISCOURSE-INSTANCE-NAME>-2025-10-01-051720-v20250919094856.tar.gz?token=1*0ceb72e0ec4c83eae2e2c76e18a0*0

I can’t reproduce this. Did you make any other changes to your instance (proxy, nginx config changes)?

1 „Gefällt mir“

hi, thanks for your quick response!

Not that I’m aware of… no, actually quite sure not. Just on September 26, 2025, I was able to successfully download the backups via link. Today, I initiated the backup while it was already updating the docker_manager in Discourse- but I don’t know if that’s somehow relevant. Otherwise, I set up the entire VPS exactly according to ‘best practice’ because I wanted to avoid causing any problems by deviating in any way. Can I provide you with any logs from the VPS or Discourse that could help analyze this? I really don’t want to have to set up the VPS again :frowning: