Would there be possible to have a separate script running that constantly monitors gpu activity?
As the miner switches between coins/algos/miners the script would have to be running separately... I'm annoyed that I don't have a way of monitoring whenever the miner is down (whether a gpu has crashed, ccminer stopped responding or the pool disconnected).
Has anyone done anything like this around here?
I use hwinfo64 + a simple batch file and powershell script to monitor and send email alerts if gpu temp or gpu power goes outside an expected range.
So if any gpu temp goes below 60 for more than a few minutes, I get an sms text with server name, sensor name, and value.
1. use hwinfo's "sensors" widget to set what to monitor and ranges to alert.
2. when the alert is triggered hwinfow will execute a command line and pass variables: the sensor name and current value
4. command line points to a batch file which calls a powershell script and passes the variables (could do this directly to powershell and not need a batch but I add server name and other stuff in batch)
5. powershell script sends an email via gmail's smtp
also for recipient address I use my phone's sms gateway address, so the alert comes as a text msg.
What setting can I change to make the auto-switch switch less frequently?
I'm seeing the miner switch and start a new instance every 1 minute. With startup time when starting a new algo, there's a few seconds lost every minute.
Can I set it not switch for say at least 10 minutes of being on a specific miner+algo?