In this case, PC was rebooted because ".*CUDA-capable.*" is counted. But you could add more error texts.
2018.07.20:22:29:33.401: GPU4 CUDART error in CudaProgram.cu:188 : unspecified launch failure (4)
2018.07.20:22:29:33.430: GPU5 CUDART error in CudaProgram.cu:188 : unspecified launch failure (4)
2018.07.20:22:29:33.454: GPU6 CUDART error in CudaProgram.cu:188 : unspecified launch failure (4)
2018.07.20:22:29:33.662: GPU3 CUDART error in CudaProgram.cu:56 : all CUDA-capable devices are busy or unavailable (46)
2018.07.20:22:29:33.662: GPU1 CUDART error in CudaProgram.cu:56 : all CUDA-capable devices are busy or unavailable (46)
2018.07.20:22:29:33.673: GPU2 CUDART error in CudaProgram.cu:56 : all CUDA-capable devices are busy or unavailable (46)
2018.07.20:22:29:34.093: GPU4 GPU4 initMiner error: unspecified launch failure
2018.07.20:22:29:34.093: GPU2 GPU2 initMiner error: all CUDA-capable devices are busy or unavailable
2018.07.20:22:29:34.093: GPU1 GPU1 initMiner error: all CUDA-capable devices are busy or unavailable
2018.07.20:22:29:34.093: GPU6 GPU6 initMiner error: unspecified launch failure
2018.07.20:22:29:34.093: GPU3 GPU3 initMiner error: all CUDA-capable devices are busy or unavailable
2018.07.20:22:29:34.093: GPU5 GPU5 initMiner error: unspecified launch failure
In this case the built in watchdog is kicked in, and by default it does this:
restart with the same command line options
2018.07.22:14:03:56.127: GPU4 CUDA error in CudaProgram.cu:329 : an illegal instruction was encountered (715)
2018.07.22:14:03:56.127: GPU4 GPU4 search error: an illegal instruction was encountered
2018.07.22:14:03:56.141: GPU1 CUDA error in CudaProgram.cu:329 : an illegal instruction was encountered (715)
2018.07.22:14:03:56.141: GPU1 GPU1 search error: an illegal instruction was encountered
2018.07.22:14:03:56.188: GPU6 CUDA error in CudaProgram.cu:329 : an illegal instruction was encountered (715)
2018.07.22:14:03:56.188: GPU6 GPU6 search error: an illegal instruction was encountered
2018.07.22:14:03:56.204: GPU2 CUDA error in CudaProgram.cu:329 : an illegal instruction was encountered (715)
2018.07.22:14:03:56.204: GPU5 CUDA error in CudaProgram.cu:329 : an illegal instruction was encountered (715)
2018.07.22:14:03:56.204: GPU2 GPU2 search error: an illegal instruction was encountered
2018.07.22:14:03:56.204: GPU5 GPU5 search error: an illegal instruction was encountered
2018.07.22:14:03:56.235: GPU3 CUDA error in CudaProgram.cu:329 : an illegal instruction was encountered (715)
2018.07.22:14:03:56.235: GPU3 GPU3 search error: an illegal instruction was encountered
2018.07.22:14:03:56.652: main Eth speed: 261.474 MH/s, shares: 1631/0/0, time: 7:22
2018.07.22:14:03:56.652: main GPUs: 1: 54.612 MH/s (347) 2: 54.556 MH/s (356) 3: 53.975 MH/s (333) 4: 32.857 MH/s (227) 5: 32.856 MH/s (196) 6: 32.617 MH/s (172)
2018.07.22:14:03:56.921: wdog Thread(s) not responding. Restarting.
I would prefer a PC reboot instead by FarmWatchBot.
I have some more errors for PhoenixMiner (3.0c).
In this case, PC was rebooted because
".*CUDA-capable.*"
is counted. But you could add more error texts.In this case the built in watchdog is kicked in, and by default it does this:
Can FarmWatchBot deal with auto miner restarts?