Implement native miner logs under the /bin/minername/log folder
UserDC-LeGrand opened this issue · 1 comments
Please implement native miner logs under the /bin/minername/logs folder, to aid with troubleshooting of past events in HiveOS. Example below shows banned events. These are logged in NVIDIA1/AMD1 logs, but erased at next miner switch.. Previous logs are therefore purged, with no way to find root cause of previous failures.
This exists by default in naitve HiveOS out of the box. I know this is possible as I see native nanominer logs in the folders below, probably triggered by custom miner arguments I used during testing ...
Alternatively, this could be implemented by maintaining a history of time-stamped NVIDIA1/AMD1 logs, rather than just the single active NVIDIA1/AMD1 log file...
This may or may not be related/associated to incident " Better Logging #395" ...
This is #395, but will leave up.
On windows side, there is no way to record screen output easily. Some miners have anti-debugging and it triggers them trying to capture output. That's why I run -log
arguments in lieu of just capturing gnu screen
output like HiveOS does.
And if I use minerlog_2022-10-21...etc. some miners will parse the -10 as a argument.
It's all so very dumb and complicated because each miner works differently. I'm trying to come up with a solution.