What Happens to Your Rig During a Power Outage? How to Set It to Restart Automatically?

Impact of Power Outages on Mining Rigs

1. Sudden power loss halts all active hashing operations immediately, interrupting block validation and transaction processing.

2. Ungraceful shutdowns risk filesystem corruption on SSDs or NVMe drives hosting the OS and wallet data.

3. GPU VRAM contents vanish without warning, potentially leaving mining software in an inconsistent state upon reboot.

4. Repeated outages accelerate capacitor wear on motherboards and PSU units, especially in rigs operating near thermal limits.

5. Pool disconnection events accumulate, lowering effective hashrate visibility and delaying reward distribution due to stale share submission.

Hardware-Level Auto-Restart Capabilities

1. Most ATX-compliant PSUs support “AC Recovery” mode, which can be enabled via BIOS/UEFI settings under Power Management.

2. Motherboards with Intel H310/B365/H410 chipsets often include “Restore on AC Power Loss” options—set to “Power On” instead of “Power Off” or “Last State”.

3. Some ASRock H series mining motherboards feature a physical jumper (e.g., PWRBTN1) that forces automatic boot after power restoration.

4. Industrial-grade PSUs like those from SeaSonic or EVGA offer configurable delay timers before re-engaging output, preventing surge-related instability.

5. Dual-PSU rigs require synchronization: secondary PSU must be configured to follow primary’s power state via 3.3V sense line coupling or relay-based triggering.

Software Configuration for Seamless Resumption

1. Linux rigs running Ubuntu Server can use systemd services with Restart=always and StartLimitIntervalSec=0 to bypass restart throttling.

2. Windows-based rigs benefit from Task Scheduler triggers set to “At system startup”, combined with batch scripts that verify driver load before launching lolMiner or T-Rex.

3. Dockerized mining containers should include health checks and auto-restart policies defined in docker-compose.yml using restart: unless-stopped.

4. Remote monitoring tools like HiveOS embed watchdog daemons that detect process absence and trigger full system reset via IPMI or BMC interfaces.

5. Custom init scripts must avoid race conditions by waiting for network interface readiness and pool connectivity before initializing miner binaries.

UPS Integration and Failover Logic

1. Pure sine wave UPS units with USB or SNMP interfaces allow rigs to execute safe shutdown sequences when battery levels drop below 30%.

2. NUT (Network UPS Tools) daemon on Linux enables conditional restart rules—only permitting boot if utility power has stabilized for at least 90 seconds.

3. Dual-circuit setups isolate mining GPUs from control logic; CPU and storage remain powered during brief brownouts while hashboards cycle independently.

4. Relay modules controlled by ESP32 microcontrollers monitor line voltage and assert reset signals only after detecting three consecutive clean AC cycles.

5. Battery-backed real-time clocks maintain accurate timestamps across outages, ensuring log integrity and proper difficulty adjustment timing.

Frequently Asked Questions

Q: Can BIOS auto-restart settings override a failed GPU initialization?A: No. BIOS-level restart only powers on the motherboard. If GPU POST fails or drivers fail to load, the system may hang at firmware level without entering OS.

Q: Does enabling auto-restart increase wear on SSDs used for mining OS?A: Not significantly. Modern NVMe drives handle thousands of power-loss events before degradation becomes measurable. Journaling filesystems like ext4 with barriers mitigate metadata damage.

Q: Will my rig reconnect to the same mining pool automatically after reboot?A: Yes—if pool URL, worker name, and password are hardcoded in the miner config file or environment variables, and no manual authentication step is required.

Q: Is it safe to enable auto-restart on rigs using overclocked memory profiles?A: Only if XMP/DOCP profiles are validated for stability across cold boots. Unstable memory training may cause repeated POST failures until manual intervention resets CMOS.