Closed StefanOberhumer closed 6 years ago
Could you attach a trace?
if you are using linux:
nohup ./grm -t > trace.txt 2>&1 &
I'll run grm with those parameters. If I catch it again I'll inform you. For privacy reasons I would prefer to send you the trace via email ?
sure. rodneymo@rigmonitor.app
@StefanOberhumer, did you ever send me that trace?
@rodneymo ... Ups - sorry - no. Have not logged any again - but I tried to figure it out. It seems while starting up the miner responses already to the API call but have not initialized the GPUs. So the GPU-list is empty or only a part. (eg: I specified 8 GPUs in rig-monitorv2 and the miner returns only info about 0 or 2)
Maybe you can check the length of the returned elements of the API call before accessing it via VAR[index]
@StefanOberhumer that should not be a problem as it doesn't expected a pre-defined number of GPUs and certainly does not assume the number of GPUs to be the number defined in the config file . Instead it counts the number of fields separated by ";". Could you send me an example?
Will try to reproduce, closing issue till I have new news.
I think one network switch in the farm got problems (all network connections to the rigs were lost):