[Bug] Connection drops after 1-2 hours while logs show "Connected" (No Reconnect Triggered)

[abolfazlakbarzadeh]

Description:

Hi,

I am experiencing an issue where the Aether connection drops silently after running for approximately 1–2 hours. Although the actual network connectivity is lost, the application logs continue to show the state as "Connected," and the service does not attempt to reconnect.

This behavior persists even with AETHER_QUICK_RECONNECT=1 enabled. The systemd service restarts the process (due to Restart=always), but the initial connection state appears to be stuck in a false-positive "connected" state until the process is killed/restarted by systemd.

Environment:

• Aether Version: 1.4.0
• OS: Ubuntu 26.04 LTS (Resolute Raccoon)
• Kernel: 7.0.0-14-generic
• Deployment: systemd service

Systemd Service Configuration:

[Unit]
Description=freedom2 (aether)
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
WorkingDirectory=/root/aether-freedom
ExecStart=/root/aether-freedom/aether

# Configuration Environment Variables
Environment="AETHER_PROTOCOL=gool"
Environment="AETHER_NOIZE=firewall"
Environment="AETHER_SCAN=turbo"
Environment="AETHER_IP=IPv4"
Environment="AETHER_QUICK_RECONNECT=1"

# Memory Management Environment Variables
Environment="MALLOC_TRIM_THRESHOLD_=131072"
Environment="MALLOC_ARENA_MAX=2"

# Resource Limits
MemoryMax=300M
CPUQuota=65%

# Restart Policies
Restart=always
RestartSec=5

# Security Hardening
NoNewPrivileges=true

[Install]
WantedBy=multi-user.target

Steps to Reproduce:

1. Install Aether v1.4.0 on Ubuntu 26.04.
2. Configure the systemd service as shown above.
3. Start the service (systemctl start aether).
4. Verify initial connection is successful (logs show wireguard tunnel validated).
5. Wait for approximately 1–2 hours.
6. Observe that network traffic stops, but logs do not show any error or reconnection attempt.
7. Check logs: The last entries remain the initial successful connection logs, with no subsequent heartbeat or error messages indicating the drop.

Expected Behavior:

• If the connection drops, Aether should detect the failure (via keep-alives or ping checks).
• The logs should reflect the disconnection.
• The service should automatically attempt to reconnect (especially since AETHER_QUICK_RECONNECT=1 is set).

Actual Behavior:

• The connection drops silently.
• Logs continue to show the initial "Connected" state from startup.
• No reconnection logic is triggered until the systemd service restarts the process (after the process crashes or is manually restarted).

Logs:

Below are the logs from the initial successful startup. Note that after ~1-2 hours, no further logs are generated until the service restarts.

Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.329Z INFO  aether] [*] establishing outer WARP tunnel to 162.159.195.1:854...
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.329Z INFO  aether] [+] aethernoize profile: firewall
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.329Z INFO  aether] [*] [outer] validating WireGuard tunnel with 162.159.195.1:854 (handshake + data-plane)...
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.530Z INFO  aether] [+] [outer] wireguard tunnel validated (end-to-end data confirmed)
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.532Z INFO  aether] [+] inner endpoint tunneled through outer warp via 127.0.0.1:36039
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.532Z INFO  aether] [*] establishing inner WARP tunnel (warp-in-warp)...
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.532Z INFO  aether] [*] [inner] validating WireGuard tunnel with 127.0.0.1:36039 (handshake + data-plane)...
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.755Z INFO  aether] [+] [inner] wireguard tunnel validated (end-to-end data confirmed)
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.755Z INFO  aether] [+] socks5 server listening on 127.0.0.1:1819
Jul 27 09:37:30 aether[709668]: [2026-07-27T06:37:30.756Z INFO  aether::socks] socks5 listening on 127.0.0.1:1819

Additional Context:

• The issue seems to be related to the keep-alive mechanism or the internal state machine not detecting the stale connection.
• I am using AETHER_NOIZE=firewall and AETHER_PROTOCOL=gool.
• The system has strict memory (300M) and CPU (65%) limits via systemd, but the process does not appear to be OOM-killed or CPU-throttled to the point of hanging (it just stops logging).

Please let me know if you need more detailed logs (e.g., with debug level enabled) or specific network traces.