This draft is here because the original troubleshooting note clearly had a real problem in it, even if the surviving details are incomplete.

The Symptoms

The FTP client was reporting TLS shutdown problems:

1
2
3

GnuTLS error -110 in gnutls_record_recv
The TLS connection was non-properly terminated
Could not read from socket: ECONNABORTED

On the Kubernetes side, the pod was also showing repeated failures and restarts.

That combination is enough to make this worth preserving as a draft even before the full root cause is reconstructed.

What the Original Note Captured

The note focused on:

  • describing the pod
  • checking logs
  • scaling the deployment
  • inspecting the container restart behavior

That suggests the failure could have been inside the FTP container itself, not just in the network path between client and service.

Why This Is Still a Draft Stub

The missing piece right now is the final diagnosis. I do not want to fake that.

Before this becomes a full post, I would want to recover or add:

  • whether the issue was certificate-related
  • whether passive FTP configuration was wrong
  • whether the container image itself was unstable
  • whether the mounted storage caused the service to crash

That said, the draft already has a useful operational shape: TLS symptom outside, crash behavior inside, and Kubernetes inspection as the bridge between the two.