System Details

• OS: Ubuntu Server

• Kernel: 5.15.0-173-generic

• Architecture: x86_64

• Setup: Headless server (accessed via SSH / Tailscale)

• Network: WiFi using a USB WiFi adapter (Realtek RTL871X driver)

Problem Summary

I have a home Ubuntu server connected to WiFi using a USB adapter.
When my home router restarts, the server:

• Disconnects from WiFi

• Attempts recovery via scripts (cron jobs)

• Sometimes recovers successfully

• But after some time, all cron jobs stop producing logs

• SSH connection drops (connection reset)

• I need to manually reboot the server to recover

Network Configuration (nmcli)

connection.autoconnect:                 yes
connection.autoconnect-priority:        10
connection.autoconnect-retries:         -1 (default)
connection.autoconnect-slaves:          -1 (default)

Cron Jobs

Running as root:

*/2 * * * * /bin/bash /home/madhur/scripts/wifi-recover.sh
*/2 * * * * sleep 20 && /bin/bash /home/madhur/scripts/ping-monitor.sh
*/2 * * * * sleep 40 && /bin/bash /home/madhur/scripts/tailscale-watchdog.sh

Scripts Overview

I have 3 scripts:

1. wifi-recover.sh

   • Checks interface state, IP, and ping

   • If down → runs:

     systemctl restart NetworkManager
     
     

   • Then brings interface up and checks connectivity

   • Restarts tailscaled after recovery

2. ping-monitor.sh

   • Pings 8.8.8.8

   • If failed → toggles WiFi using:

     nmcli radio wifi off/on
     
     

3. tailscale-watchdog.sh

   • Checks Tailscale health via JSON

   • Restarts tailscaled if offline

Observed Logs

WiFi Recover

[12:52:04] Wi-Fi DOWN → Restarting NetworkManager...
[12:52:45] ❌ Recovery FAILED (No IP assigned)

[12:54:01] Wi-Fi DOWN → Restarting NetworkManager...
[12:54:41] ✅ Recovery SUCCESS (IP assigned)

Ping Monitor

[12:52:04] PING FAIL → 100% packet loss
[12:54:01] PING FAIL → Network unreachable
[01:16:03] PING OK

Tailscale

[12:54:01] Online=false → restarting tailscaled
[12:54:09] ✅ restarted

Critical Issue

After the router restart (~12:54 AM):

• Recovery appears successful

• But:

  • SSH disconnects (connection reset)

  • Cron logs stop appearing after that point

  • Scripts no longer seem to execute

• System becomes unreachable until manual reboot

Additional Observations

From journalctl:

• NetworkManager is restarted multiple times

• WiFi driver logs:

  • RTL871X: nolinked power save

  • firmware reloads

• Tailscale shows DNS failures:

  • Temporary failure in name resolution

Questions

1. Why do cron jobs stop producing logs after network recovery?

   • Is cron actually stopping, or are scripts failing silently?

2. Is restarting NetworkManager from cron a bad practice?

   • Could this be causing system instability or race conditions?

3. Could the USB WiFi adapter (Realtek RTL871X) be causing instability?

   • Known issues with power saving / driver resets?

4. What is the correct way to ensure automatic WiFi reconnection on Ubuntu Server?

5. Should I:

   • Avoid systemctl restart NetworkManager?

   • Use only nmcli commands?

   • Use a systemd service instead of cron?

6. How can I make the system:

   • Recover reliably after router restart

   • Keep cron/scripts running

   • Avoid needing manual reboot

Goal

I want a fully autonomous recovery system such that:

• If WiFi/router restarts → server reconnects automatically

• Tailscale reconnects

• SSH remains stable (or recovers)

• No manual intervention required