problems with creating binary on custom s3 instance about fcrepo HOT 4 CLOSED

@BlackLotus : A few questions: 1) are you able to reliably reproduce the issue? 2) have you tried this against S3 proper? Just looking at the stacktrace, it appears that there was a stream error in transferring between the s3 client and the s3 conforming service. That is the root issue. A secondary issue appears to be that the jcr clean up in the wake of the failure didn't occur for some reason.

It would be helpful if you would create an issue on https://jira.duraspace.org/ with a clear recipe for reproducing the error and against an environment that we could reasonably reproduce (ie against AWS s3.).

from fcrepo.

Ok sorry for the delayed response, but since I didn't use aws I didn't think I could reproduce the problem with aws. So here is what I did.

First I got an aws account and simply tried fcrepo with it - it worked.
Next I tried to find out what amazon aws was doing different than ceph for this I configured an nginx reverse proxy and set up my local dns to forward every request to 127.0.0.1 instead of the amazon server. My expectation was that I could easily monitor the traffic going to amazon and the traffic going to ceph and spot either the difference in behaviour or the offending request that didn't work with ceph. What I got instead was the same problem with amazon s3. I tried to mimic amazons s3 behaviour as good as I could and after failing I tried to follow https://aws.amazon.com/de/blogs/compute/nginx-reverse-proxy-sidecar-container-on-amazon-ecs/
Anyway this is one of my nginx configs I came up with

  gzip_proxied any;
  gzip_types text/plain application/json;
  gzip_min_length 1000;

server {
    listen 80;
    server_name .s3-proxy .s3.us-east-1.amazonaws.com;
    client_max_body_size 10g;

    access_log /var/log/nginx/s3proxy-access.log postdata;
    error_log /var/log/nginx/s3proxy-error.log;


    proxy_set_header Host $host;
    proxy_buffering off;

    location / {
        fastcgi_pass_header Authorization;
        fastcgi_pass_request_headers on;
        include fastcgi_params;
	proxy_pass https://52.216.101.21;
      proxy_pass_header Server;
      proxy_http_version 1.1;
      proxy_set_header Upgrade $http_upgrade;
      proxy_set_header Connection 'upgrade';
      proxy_set_header Host $host;
      proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
      proxy_cache_bypass $http_upgrade;
    }
}

I also tried other reverse proxy solutions without success. I will later create an issue at jra.duraspace.org, but I would have really liked if I understood why it only seems to work with the original s3.

from fcrepo.

I thought the custom endpoint flag was not allowed where can I set it, I would like to test this as well but with min.io

from fcrepo.

Fedora 5 is no longer in active development as Fedora 6 is now the LTS version of Fedora. It looks like the Samvera/Valkerye development was completed to support this though. Closing this here now

from fcrepo.