Nexardo

Satisfactory Field guide

Satisfactory «Connection to Host Lost» Fix: Timeouts, MTU, and Network Config

Diagnosing and fixing the "Connection to the host has been lost" error on Satisfactory dedicated servers: Engine.ini timeout tweaks, MTU packet fragmentation fixes, and UDP socket configuration.

EugenePublishedUpdatedReading time4 min
Satisfactory «Connection to Host Lost» Fix: Timeouts, MTU, and Network Config

The error "Connection to the host has been lost" is one of the most frustrating network interruptions encountered on Satisfactory dedicated servers. The disconnect typically triggers at two critical operational moments: during the initial join sequence while parsing massive factory states, or during periodic autosaves when the server temporarily freezes while writing world archives to disk. The game client misses network heartbeats beyond its configured timeout threshold and aborts the session.

This guide examines Unreal Engine NetDriver mechanics within Satisfactory, details critical timeout adjustments in Engine.ini, resolves MTU packet fragmentation bottlenecks, and tunes autosave cycles to guarantee uninterrupted multiplayer sessions.

Why Satisfactory Dedicated Servers Drop Player Connections#

The Satisfactory dedicated server operates over UDP protocol on port 7777 utilizing Unreal Engine's native NetDriver. Unlike simple arena games, a late-stage Satisfactory factory simulates hundreds of thousands of coordinated actors: conveyor belts, splitters, pipeline networks, railway engines, and automated manufacturing machines.

Network disconnections trace back to three primary culprits:

  1. Network Timeouts During Autosaves: Serializing a 50–100 MB factory state on slower drives halts the primary server thread for 10 to 30 seconds. If ConnectionTimeout is shorter than this I/O freeze, the client presumes the server crashed and drops back to the main menu.
  2. MTU Packet Fragmentation: Connecting players must download an initial factory snapshot. If the UDP datagram size exceeds network MTU (standard 1500 bytes) and intermediate routers discard fragmented packets, connection handshakes stall permanently on the "Joining" screen.
  3. Client Bandwidth Throttling: Satisfactory ships with a default network quality preset that artificially caps client data ingestion.

Tuning Network Timeouts in Engine.ini#

To grant the engine sufficient headroom during disk writes and heavy scene loading, modify the configuration on your Linux host: /opt/satisfactory/FactoryGame/Saved/Config/LinuxServer/Engine.ini

Stop the dedicated server service before editing configuration files:

bash
sudo systemctl stop satisfactory

Append or update the IpNetDriver section at the end of Engine.ini:

ini
[/Script/OnlineSubsystemUtils.IpNetDriver]
ConnectionTimeout=90.0
InitialConnectTimeout=180.0
NetServerMaxTickRate=30
LanServerMaxTickRate=30
MaxNetTickRate=30
  • ConnectionTimeout=90.0 — extends in-game heartbeat tolerance to 90 seconds (up from the default 30 seconds), preventing disconnects during large autosaves.
  • InitialConnectTimeout=180.0 — provides up to 3 minutes for clients on high-latency connections to ingest complex initial factory snapshots.
  • NetServerMaxTickRate=30 — stabilizes network updates at 30 Hz, optimizing CPU utilization without compromising synchronization fidelity.

Mitigating MTU Packet Fragmentation and Interface Diagnostics#

If players experience drops despite extended timeouts, examine the Maximum Transmission Unit (MTU) on your server's primary network adapter:

bash
ip link show | grep mtu

Execute a ping test with the Don't Fragment (DF) flag enabled to identify MTU bottlenecks:

bash
ping -M do -s 1472 -c 4 1.1.1.1

If the probe returns Frag needed and DF set, an upstream tunnel or hosting provider utilizes reduced MTU ceilings. Lower your network interface MTU to 1420 or 1400 accordingly:

bash
sudo ip link set dev eth0 mtu 1420

Verify active UDP listening sockets using ss:

bash
sudo ss -ulnp | grep 7777

Adjusting Autosave Intervals to Prevent Server Freezes#

The factory default autosave interval in Satisfactory is 5 minutes (300 seconds). On sprawling megabases, saving every 5 minutes introduces constant stutter and elevated disconnect vulnerability.

Lengthen the autosave interval to 20 or 30 minutes in /opt/satisfactory/FactoryGame/Saved/Config/LinuxServer/GameUserSettings.ini:

ini
[/Script/FactoryGame.FGSaveSession]
mAutosaveInterval=1200

Setting 1200 seconds enforces a 20-minute cycle, minimizing unnecessary world serialization overhead.

Through a modern server management panel, dedicated NVMe storage tiers and network buffer optimizations eliminate micro-stutters during state flushes.

Configuring Client-Side Network Settings#

In addition to server configuration, joining players should optimize client-side network limits:

  1. Open the game menu: Options → Gameplay.
  2. Locate the Network Quality dropdown.
  3. Select Ultra.

The Ultra setting removes the legacy 30 KB/s client bandwidth limiter, allowing the client to ingest dense factory updates without overflowing local network buffers.

Start the Satisfactory daemon and verify socket binding:

bash
sudo systemctl start satisfactory
sudo journalctl -u satisfactory -n 50 --no-pager | grep -i "netdriver"

With these comprehensive network parameters applied, the "Connection to the host has been lost" error is resolved even across complex multi-player industrial bases.

Sources#