We are aware images aren’t loading and new images can’t be uploaded. It appears to be a bug with the latest Lemmy update as we’re not the only instance with this problem. We have filed a bug report and are looking into possible solutions and hopefully it will get fixed soon. We are also looking into better ways to communicate server issues like this. Thank you for your understanding! We are learning this alongside all of you.
EDIT: If anyone out there has an idea to the cause please feel free to shoot me a DM!
EDIT 2: We’re back in action folks! Thanks for your patience, you may feel free to set avatars now.
Hi, armchair dev shooting from the hip here. Glancing at the error and docs, it seems like adding an entry in /etc/hosts aliasing “pictrs” to “localhost” on the host or container where lemmy is running might help.
Thanks for letting us know.
I wasn’t sure if you’d blocked direct posting of photos and images for reasons, or if it was part of the larger issue that’s affecting the headers.
Unfortunately, I’m still seeing the same errors when trying to load my very small-sized avatar. Not fixed yet for me.
@crashdoom@pawb.social I know you fixed this on our instance, any idea to help the lads here?
We’re using the Ansible playbook deployment, and I ended up giving the
pictrsservice a restart through docker (docker restart <id>and you can get the id by usingdocker ps).It didn’t seem to be out of space or even offline, it just locked up and stopped responding to both new uploads and existing image requests.
Our pictrs service seems stuck in a restart loop.
Not sure what kind of storage is backing it, but if you’re using filesystem storage, the pictrs data directory and files need to be chown’d to 991:991 and/or match the uid/gid in the user line of the docker compose. 991:991 is what’s in the documentation
