HiveOS 100% unlock bug, fan drop to zero and hashrate drop to very low value

[linlexuan2005]

Fan drops to 0 and hashrate drops to about 1/3 of full hashrate. Restarting miner cannot solve it, have to restart rig.

Hvieos挖了一会，会出现一张显卡风扇转速变为0，算力降到一二十M，重启miner无果，只能重启系统恢复，不知道这跟nbminer有没有关系

[NucleaR-II]

Hello,

with me unfortunately also after some time a card shows 0% fan and the MHS go down completely.

Hive OS: 0.6-217@220505

5.4.0-hiveos #140 Nvidia Driver: 510.60.02 nbminer v.41.0

no matter if i turn autofan on or off

does anyone have any idea what i can do ?

THANK YOU VERY MUCH NBMINER !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[Nitin00700]

Same issue

[Nitin00700]

Please someone help us My oc are correct and iam also using recommended drivers by NB Miner

[ghost]

same issue here and rig need to be restarted every one hour or so < Thanks NBminer and plz fix

[meshersky]

I have this error in log:

[19:48:59] INFO - ethash - New job: ru-eth.hiveon.net:4444, ID: a9eeb4b1, DIFF: 4.295G
[19:49:00] ERROR - CUDA Error: unspecified launch failure (err_no=4)
[19:49:00] ERROR - Device 2 exception, exit ...
[19:49:01] ERROR - !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
[19:49:01] ERROR - Mining program unexpected exit.
[19:49:01] ERROR - Code: 6, Reason: Process crashed
[19:49:01] ERROR - Restart miner after 10 secs ...
[19:49:01] ERROR - !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

And in syslog:

May  8 19:48:52 hive3080_ hive-watchdog[1051]: OK LA(5m): 1.38 < 22.0, LA(1m): 1.21 < 44.0
May  8 19:49:00 hive3080_ kernel: [ 2027.507045][  T921] NVRM: GPU at PCI:0000:03:00: GPU-1b7dc474-13c3-3bf3-6fff-5b1a7fe54313
May  8 19:49:00 hive3080_ kernel: [ 2027.507048][  T921] NVRM: Xid (PCI:0000:03:00): 62, pid=921, 0000(0000) 00000000 00000000
May  8 19:49:00 hive3080_ kernel: [ 2027.550062][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000010
May  8 19:49:00 hive3080_ kernel: [ 2027.555690][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000011
May  8 19:49:00 hive3080_ kernel: [ 2027.556731][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000012
May  8 19:49:00 hive3080_ kernel: [ 2027.557676][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000013
May  8 19:49:00 hive3080_ kernel: [ 2027.558631][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000014
May  8 19:49:00 hive3080_ kernel: [ 2027.559575][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000015
May  8 19:49:00 hive3080_ kernel: [ 2027.560519][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000016
May  8 19:49:00 hive3080_ kernel: [ 2027.561466][  T921] NVRM: Xid (PCI:0000:03:00): 45, pid=4761, Ch 00000017

[NucleaR-II]

[dk7988]

i was having odd hashrate problems too.. i updated to driver 510.68.02 and my problems are fixed now

[meshersky]

i was having odd hashrate problems too.. i updated to driver 510.68.02 and my problems are fixed now

My problem was not solved in this way.

[dk7988]

is it a mixed rig? with amd and nvidia cards?

[meshersky]

is it a mixed rig? with amd and nvidia cards?

not, only 3080ti, msi gx trio and gigabyte

[tomcruise946]

How to update to 510.68.02

On Mon, 9 May, 2022, 12:40 am dk7988, @.***> wrote:

i was having odd hashrate problems too.. i updated to driver 510.68.02 and my problems are fixed now

— Reply to this email directly, view it on GitHub https://github.com/NebuTech/NBMiner/issues/818#issuecomment-1120471563, or unsubscribe https://github.com/notifications/unsubscribe-auth/AURKUGUEDO7ZK5DOXV52EJDVJAGS3ANCNFSM5VMDNIFA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[tomcruise946]

How can I upgrade to 510.68.02. All cards are crashing so have changed it back to trex miner 49 mhs again.

[dk7988]

• meshersky: what hiveos version are you using? i updated to 0.6-217@220505

• tomcruise946: assuming your using hiveos, you have some options 1.) re-downlond and re-flash hiveos on a new usb/hard drive 2.) remote on to the miner and use command "nvidia-driver-updated" (note: this will default to the newest nvidia drive on hive) to pick another drive to download and install you can use command "nvidia-driver-updated --list" 3.) you might be about to use "Worker Command" too, in the hiveos worker/miner webpage in the upper most/ top most banner look for and click on the button that looks like ">_" (to use worker commands)

[meshersky]

• meshersky: what hiveos version are you using? i updated to 0.6-217@220505

Last version with last drivers. I'll try to downgrade drivers to 510.60.02

[dk7988]

in an attempt to be helpful here is everything i had to update to for my lhr cards to work correctly

i also had a hell of a time trying to download and install the updated drivers from hive's archive location i had to do some hoopy things in order for the install to completely work right. everything was reporting the driver was updated but when i would remoted onto the miner all my nvidia cards said "Malfunction"

if any one is also running into this problem (with the updated drivers not installing correctly) i found and followed this https://medium.com/@overcookedpanda/installing-the-latest-nvidia-drivers-on-hiveos-fe302f214570

i did deviated from this though.. (note: i had completely downloaded the updated driver from hive's archive by remoting onto the miner then using hive's command "nvidia-driver-updated" . Halfway through the install at the step where it says something about dwk_something blah blah is where it would fail then attempt to restart the nvidia service and it would fail every time even after rebooting)

i ran the following commands after the download completed and while remoted on the miner:

miner stop killall xinit sudo su sudo init 3 mkdir /temp export TMPDIR=/temp cd /hive-drivers-pack/ ls ##<<--- at this point a list shows all the downloaded drivers i saw driver "NVIDIA-Linux-x86_64-510.68.02.run" was an option chmod +x NVIDIA-Linux-x86_64-510.68.02.run ./NVIDIA-Linux-x86_64-510.68.02.run

then the install process started i used all the default selections for all the following questions and after waiting a little bit everything took and the driver was now in fact installed then ran command

sudo apt-get update; sudo apt-get install nvidia-settings -y

then rebooted and everything (lhr cards were running at full hash and fans were at targeted speed) started working as intended

[AshkanAbd]

I have tested all 510.xx.xx drivers. but there is no fix yet.

[brasrox]

same here, 4 rxt 3070ti with 900/2000 clock no pl ms down to 30 and fan goes to 0 only back mining if i reboot rig using 510.68.02 here hiveos.

reduced mem to 1900 and testing

[ansonk4]

I have the same issue using 510.60.02

[tomcruise946]

Cant abel to download 510.68.02 drivers

[l-Kage-l]

Same issue for me too. 2Rigs running only 3060Tis, happens every 30 mins to 1hour and rig needs a restart running on HiveOS with 510.60.02 drivers. No overclock change has worked towards resolving the problem. Looking forward to a fix

What is really strange is that my 3080Ti has no problem with this yet, running for more than 8h straight with no problem.

Seems to be a Linux only problem from reading on some forums.

Update: 3080Ti crashed too after 10 hours

[chechle39]

I have the same issue

[AshkanAbd]

Guys anyone use windows for mining and got this issue? Also anyone with 8GB of RAM has this issue?

[NucleaR-II]

Guys anyone use windows for mining and got this issue? Also anyone with 8GB of RAM has this issue?

have only HIVEOS , unfortunately can not change it so quickly to test it

[Gkozd]

"Note2: If you run into issues, please change driver to recommened versions, and set your memclock 100 ~ 200mhz lower that previous LHR partial unlock situation" Or even ~300mhz lower. Still, spectacular results. Do not be greedy too much for now. Set up watch-dog in HiveOS to restart entire rig, if hashrate has dropped too much (prevent from cooking your g-card's).

[AshkanAbd]

I think this issue happens only in linux so maybe because of swap?!!

[aragonense]

on hiveos: try set PL to max bios limit

[Tommy742]

apparently solved for my 3060s v2 with 510 drivers and 1552 core 2200 mem pl 120 48.3mhn on hiveos stable for the while. run only my V2 cards on it if rig is mixed. and look for some less agressive OC than before that's how i fixed it. go lower on mem till catch stability dont be to greedy ; )

[claremoresigns]

apparently solved for my 3060s v2 with 510 drivers and 1552 core 2200 mem pl 120 48.3mhn on hiveos stable for the while. run only my V2 cards on it if rig is mixed. and look for some less agressive OC than before that's how i fixed it. go lower on mem till catch stability dont be to greedy ; )

interesting, I too lowered my 3060 and 3060ti mem clocks and it seemed to last longer but still 0 fan and 1/3 hashrate after a bit....

[iKonTechDev]

I had this going on i have been ok for about 4 hours so far on all the rigs try driver: 510.68.02 Use command: nvidia-update-driver https://us.download.nvidia.com/XFree86/Linux-x86_64/510.60.02/NVIDIA-Linux-x86_64-510.60.02.run

Seemed to work for me i tried clock and all that stuff nothing seemed to get me past an hour but this here! GL

[mohsenk94]

I had this going on i have been ok for about 4 hours so far on all the rigs try driver: 510.68.02 Use command: nvidia-update-driver https://us.download.nvidia.com/XFree86/Linux-x86_64/510.60.02/NVIDIA-Linux-x86_64-510.60.02.run

Seemed to work for me i tried clock and all that stuff nothing seemed to get me past an hour but this here! GL

I have this driver... Does not work GPU fans go to 0 one by one over time and hashrate locks at 30% of full hash...