Hi fullzero, Hi Guys,
(Long time i was in r/o mode. This is my first post.)
What you think about implementing into NVOC nvidia-docker and run miner for each isolated GPU in docker container?
It will be more fail-safe for rig with many GPU`s, more flexible to individual OC and give a way to launch several different miners.
for example: One of my rig have 13 GPU different models => OC settings are different for each others.
If one GPU crash then crash all GPU`s and whatchdog restart miner. In my case restartin 3main takes wery long time, actually wery much time takes running nvidia-settings.
I was forced to set in sleep timeout in wdog equal to 120s otherwise i have loop when wdog restarts 3mian.
p.s. exuse me for poor english
(Long time i was in r/o mode. This is my first post.)
What you think about implementing into NVOC nvidia-docker and run miner for each isolated GPU in docker container?
It will be more fail-safe for rig with many GPU`s, more flexible to individual OC and give a way to launch several different miners.
for example: One of my rig have 13 GPU different models => OC settings are different for each others.
If one GPU crash then crash all GPU`s and whatchdog restart miner. In my case restartin 3main takes wery long time, actually wery much time takes running nvidia-settings.
I was forced to set in sleep timeout in wdog equal to 120s otherwise i have loop when wdog restarts 3mian.
p.s. exuse me for poor english
I would have to look at how well the docker containers interface with the hardware. I have used docker containers before; but I'm not sure if this is viable; my guess is it is not.