A miner that stops at 2am and stays stopped until you notice the next morning has wasted hours of work. A watchdog fixes that: it watches the miner process and restarts it the moment it dies. This guide covers a simple, dependable setup on both Windows and Linux.
Why miners stop in the first place
The usual culprits are network blips, a power-saving sleep event, a driver hiccup, or an overnight router reboot. Most are transient. The miner could happily continue if something simply relaunched it. That something is your watchdog.
Windows option 1: a restart loop script
Create a small batch file that launches the miner inside a loop. When the miner exits for any reason, the loop waits a few seconds and starts it again:
- Write
:loop, then the line that runs your miner executable, thentimeout /t 10, thengoto loop. - Save it as
run-miner.batand launch the miner through this file instead of directly.
Windows option 2: Task Scheduler at login
Open Task Scheduler, create a task that runs your run-miner.bat at user logon, and enable the option to restart the task if it fails. This also handles the case where the whole PC reboots - the miner comes back automatically once you are logged in.
Linux option: a systemd service
Systemd has restart-on-failure built in. Create a unit file that runs your miner, set Restart=always and a short RestartSec=10, then enable the service so it starts on boot. Systemd will relaunch the miner after crashes and after reboots without any scripting.
Add a hashrate watchdog (optional but smart)
A process can be alive but stuck producing zero hashrate. Many miners support an internal watchdog setting that restarts the miner if hashrate drops to zero for a set period. Enable it in the config if available. If not, a short script that checks the log for recent share accepted lines and kills a silent miner achieves the same thing.
Test it before you trust it
- Manually kill the miner and confirm it relaunches within your timeout.
- Reboot the PC and confirm the miner returns on its own.
- Unplug the network briefly and confirm it reconnects or restarts cleanly.
Keep it quiet but logged
Redirect the miner output to a log file so you can review what happened overnight. A watchdog that restarts silently is good; one that also leaves a trail is better for spotting a recurring problem.