hi, all!
I'm using nvOC for a while and today got some rigs broken, tried to see what's wrong and got this error when launching nvidia-smi: "Failed to initialize NVML: Driver/library version mismatch". This happened to every nvidia rig that I tried to reboot. Never touched the drivers in nvOC distro, so I'm confused how this could happen... Does anyone have ideas? Is there any kind of autoupdater which could change something after reboot? How to disable it and make everything working again?
This exactly happened to me this morning!!
Don't know what has triggered this! haven't rebooted my for almost 2 weeks as it was running flawless!! I have restarted it, can connect through teamviewer but cannot see anything, just some blank screen!!
It took 3 hours to get my RIG back UP with version 19.13
My last telegram message ended up like this
System Boot Time: 2017-10-11 05:11:19
System Up Time: up 1 week, 6 days, 57 minutes
Miner Uptime: 16:13:22
GPU Count: 1
GPU_UTILIZATIONS: Failed to initialize NVML: Driver/library version mismatch
TEMPS: Failed to initialize NVML: Driver/library version mismatch
POWERDRAW: Failed to initialize NVML: Driver/library version mismatch
FAN SPEEDS: Failed to initialize NVML: Driver/library version mismatch
No idea why it happened, can't even access logs because of no access.
I had 3 wallets in there, I need to access it soon or rather than later! (haven't made any backup for them).
I have the same issue. Both of my rigs running nvOC are down. It looks like some kind of update was done and drivers stopped working.
This is all of a sudden asking me a login password on both my rigs and when i input miner1 it does nothing.
Nvidia driver updated today on one of my rigs
A reboot fixed my problem
The other one already had latest 384-90
run update / upgrade again
sudo apt update && sudo apt upgrade && sudo reboot
So for those who are now reaching the login screen and cannot login even when you type in your password. SSH into your rig and run the following commands:
sudo apt-get purge nvidia*
sudo dpkg --configure -a
sudo shutdown -r now
Now everything's working back to normal. Not sure why it did that for me. Also, I don't know if the first commmand is necessary, it's just what I typed in initially so I included it.
Too bad I don't have ssh configured/enabled.
If this happens you should be able to use console mode to fix the botched update. At the login screen:
press:
ctrl + alt + f1
to enter console mode
login is:
m1
password:
miner1
then use:
sudo dpkg --configure -a
then:
sudo apt update && sudo apt upgrade && sudo reboot
Nvidia driver updated today to version 384.111. I did as same as above. Everythings work fine, except power usage got ERR! in nvidia-smi and in miner shows 0w. I guess because of the new driver.
I'm wondering if there's any way I can downgrade Nvidia driver to 384.90 which is my old driver version. If it's possible please teach me step by step, I'm very new to Linux. Thank you very much.