Rudder

While testing with Rudder Version 6.2.4 I noticed the following problem.

I used File download (Rudder server) to copy a large file (~1GB) to a node (each time slightly modified by appending a single character).
The files in /var/rudder/modified-files/ eventually filled up the filesystem. (as expected)
When I then try again to copy the file with rudder to the filed up filesystem of the node, the execution of rudder agent run takes very long (in my case ~1000 seconds) and needs 100% CPU.

rudder agent run terminates at some point, but there are still cf-agent processes left over that continue to use 100% CPU.

2021-04-18T15:58:10+00:00    error: Failed to write to destination file (write: No space left on device)
2021-04-18T15:58:10+00:00    error: Local disk write failed copying '<SERVER_IP>:/var/rudder/configuration-repository/shared-files/file2.txt' to '/root/file2.txt.cfnew'
2021-04-18T16:13:09+00:00    error: Was not able to copy '/var/rudder/configuration-repository/shared-files/file2.txt' to '/root/file2.txt'
907.43s E| error         copyFile                  Copy file                 /root/file2.txt    The content or permissions of the file(s) could not have been repaired (file file2.txt not found?)
  0.02s E| n/a           copyFile                  Post-modification hook    /root/file2.txt    No post-hook command for copy of file2.txt to /root/file2.txt was defined, not executing
2021-04-18T16:13:10+00:00    error: SSL read after retries: underlying network error ()