Exploding Heads - Help hold the ruling class accountable
exploding-heads.com

I ran out of space on my instance which is hosted on linode. I used their automated system to upgrade the node so it had more hard drive space.

Since I did that, images have not displayed on my instance. Either from existing posts or new posts. How do I fix this?

I submited a question on GitHub https://github.com/LemmyNet/lemmy/issues/2637 and was asked to provide a “the docker compose logs” and to try restarting “docker compose”.

I rebooted the virtual server, which I assume would restart the docker compose. If not, how do I restart it?

Also can you give me the exact commands to:

  • do a docker compose restart and where I do it from
  • generate the docker compose logs and how to limit the size if it is likely to be to big to upload to github and where to run the commands from.

Thanks in advance for your help.

  • poVoq@slrpnk.net
    4·
    2 years ago

    With a SSH terminal go to the directory where the specific docker-compose.yml file is and do this:

    https://docs.docker.com/engine/reference/commandline/compose_up/

    Just restarting the VM or the containers does not recreate them. Be sure all the persistent data is properly stored and backed up as recreating the container will overwrite the data otherwise.

    But are you sure you want to run a public service with Docker if your level of understanding of it (and online search skills) is so low?

    • letstryOP
      1·
      2 years ago

      Thanks

      I did the docker-compose restart. It is still not working.

      So I went to post a test image so the logs would show what happens and I got this error in the UI

      SyntaxError: Unexpected token ‘R’, “Request er”… is not valid JSON

      • poVoq@slrpnk.net
        1·
        2 years ago

        Try the force-recreate option to make sure you are not only restarting the containers but actually recreating them.

        • letstryOP
          1·
          2 years ago

          I ran the force-recreate option.

          Now when I go to the instance I get a spinning wheel where the posts would normally appeal. I am assuming this is just rebuilding all the pages and may take some time? Would take be correct?

          • poVoq@slrpnk.net
            1·
            2 years ago

            no that looks more like it has issues connecting to the database. Maybe the database container got renamed?

            • letstryOP
              1·
              2 years ago

              so what should I do now?

              Should I restore the backup from before I ran the recreate command or do something else?

  • skariko
    2·
    2 years ago

    Hello, I have the exact same problem as you, and it also happened to me at the time when my instance was running out of free space.

    Did you by any chance solve it somehow?

      • DessalinesMA
        1·
        2 years ago
        pictrs_1    | Error: Error in DB, Read corrupted data at file offset None backtrace ()
pictrs_1    | 
pictrs_1    | Caused by:
pictrs_1    |     Read corrupted data at file offset None backtrace ()

        Its a pictrs issue, not with lemmy. @asonix@lemmy.ml or ask in the pict-rs matrix channel.

        • letstryOP
          1·
          2 years ago

          I still have not been able to fix this problem.

          I see V0.17 makes some changes to images - would the upgrade solve the problem for me?

          If not, should I set up a new instance and export the database from my existing instance and import into the new? If so, any tips on how I would do this?

          Any other suggestions?

          • letstryOP
            1·
            2 years ago

            Good news - I fixed this issue prior to the upgrade