Error connection_closed

The server closed the connection

A machine answered, then hung up before finishing. Usually the NFS service is restarting, a TCP wrapper rule rejected you, or the NAS went to sleep mid-conversation.

What this means

The TCP connection was established — so the address and port are right — but the server ended it without replying. That’s different from a refusal (nothing listening) or a timeout (nothing answering).

Common causes

• The NFS service is restarting on the NAS (after a settings change, an update, or a nightly task). Wait thirty seconds and try again.
• TCP wrappers — on Linux servers, /etc/hosts.deny with a line like rpcbind: ALL or mountd: ALL makes the server accept the connection and immediately close it. Add your subnet to /etc/hosts.allow, or remove the deny rule.
• The NAS went to sleep partway through. Disable HDD hibernation for testing, or turn on wake-on-LAN and wake it first.
• A proxy or “smart” firewall in the path is terminating idle connections. Rare on a home network, common on a corporate one.
• A userspace NFS server (Docker images, small Go/Rust servers) that crashed on the request. Check its log.

What to try

Tap Inspect Server on the server card. If the RPC table loads but the export list doesn’t, the problem is with mountd; if nothing loads, the whole RPC stack on the server is refusing you.