I’ve got a site that’s notifying an admin of backup failure, but the backup succeeds. Here’s the logs:
[2025-09-23 09:40:31] Removing tmp '/var/www/discourse/tmp/backups/default/2025-09-23-093005' directory...
[2025-09-23 09:40:31] Uploading archive...
[2025-09-23 09:41:37] EXCEPTION: Sidekiq::Shutdown
[2025-09-23 09:41:37] /var/www/discourse/vendor/bundle/ruby/3.3.0/gems/aws-sdk-s3-1.182.0/lib/aws-sdk-s3/multipart_file_uploader.rb:199:in `value'
/var/www/discourse/vendor/bundle/ruby/3.3.0/gems/aws-sdk-s3-1.182.0/lib/aws-sdk-s3/multipart_file_uploader.rb:199:in `map'
/var/www/discourse/vendor/bundle/ruby/3.3.0/gems/aws-sdk-s3-1.182.0/lib/aws-sdk-s3/multipart_file_uploader.rb:199:in
So it looks like sidekiq is restarting, so it generates the error, but it’s caught, so the backup actually succeeds. I’m guessing (withtout looking at commits) that this is an improvement over the previous behavior when these backups would get left on local storage and eventually fill up the disk.
I’m going to increase memory for sidekiq next time the container gets rebuilt.