So can gitlab runner create a non-existing s3 bucket?
When reading the documentation I thought it can with the provided credentials. In my case I use role assigned to the ec2 instance with the Gitlab runner and in the cache.s3 section of the config.toml I specify path and bucket name.
But if I do not create that bucket manually in the AWS S3 console-> gitlab job ( I use this one Get started with GitLab CI/CD | GitLab for testing + added caching section) will say uploading cache zip 404 path not found.
When I create that bucket manually everything works.