[OS] nvOC easy-to-use Linux Nvidia Mining

[fk1]

@fullzero: I'll mine some time ZEC for you since I dont have wallets yet but interested in what my small rig is capable of in terms of sol/h. mining into the placed config now as worker 'fk1bitcointalk'. cheers!

[fullzero]

All once again good afternoon!

I added:
1. email alarms (sendemail),
2. webUI for monitoring (apache2 + go + gotty),  
# All needed apts will install in first boot if $GOTTY="YES" || $EMAIL="YES". in settings.sh
3. auto update pasteBin and restart miner if something changes.
# you need to run crontab -e and add "*/5 * * * * ~/update_wallets.sh" . To check update every 5 minutes.


Completed / redone:
1. Option to set all Target temps and PowerLimits in two strings while using autoFANS
2. add specific PL and CLOCK for some coins
3. output nvidia-smi to screen every 30 sec when check fans, even if the fans are in manual mode. (to see temps, RPMs, memory in gotty, or via SSH, or via local quake)
4. Manual setting WORKER_NAME to use with pools. In settings.sh . If don't set using nvXXX (last digs of ip)
5. WORKER_NAME is hostname of rig, for more useful network search.

I divided the code into several files (start, miner, overclock, e-mail, fans). Settings rendered in 3 separate files:
settings.sh - constant machine settings
gpus_settings.sh - constant settings of video cards
wallets - settings of wallets, miners, machine settings variables.
Unfortunately, I have not yet had a chance to check how it works on the v18, but on v17 all is good. I think I can check in the next couple of days.

All the files are added to the folders as in the archive. Your need to change autostart from "gnome-terminal" to "~/start" (sudo nano ~/.config/autostart/webmob.desktop )
Settings are given in three files in the mnt sub folder. settings, gpus_settings, wallets

I would be happy if someone at least a small part of my work is useful.

fullzero and other developers: I ask for applications for deleted comments and descriptions, maybe even in some places copyright. While working on the code - some parts were in the way.

Of course, I could forget or confuse something, but I seem to remember everything and wrote.

Please do not make alternate releases.

As I have already told you:

https://bitcointalk.org/index.php?topic=1854250.msg20058141#msg20058141

In general I would like to offer as many options to members as possible and encourage members to share their customizations.

With any contribution, so long as it is not a blatant copy of a prior contribution; I will integrate it and add acknowledgement and BTC donation address to the top of oneBash.

If you use or make an executable; you must provide me with the source code and I will examine and compile it myself.

Your contribution must be free and open source.

I would like to avoid solutions involving the installation of multiple programs or services; to so something that can be done via SSH, SFTP, or a socket.

also see:

https://bitcointalk.org/index.php?topic=1854250.msg20093039#msg20093039

You can make modules to do distinct operations: like overclocking, miner selection, watchdog, monero mining, etc to make easier to tune/restart some parts.  And I will integrate them with switches to be readinto / not used in oneBash.  In addition they could also be used independently, such as a OC bash could be run while already mining.  A benchmarking bash / py could be made to find and log the opt stable settings ect.  The goal is to provide members with as many options as possible while maintaining ease of use.

Once most of the requested features have been implemented; I will make a github and target oneBash optimization.

It makes no sense to have members downloading multiple versions of the OS when only a few files are changed.  

Also; in such cases: I cannot confirm that the system has not been altered or malware added ( I am not saying you have done this, but that it is always a possibility especially when done by new members).  


If you post the code for each additional file (or send me a pm with it); I will integrate your contributions into 1bash (or selectively in 2unix depending on how different they are) with a single or multiple switches: and add acknowledgement and your BTC donation address to the top of 1bash.

I was not going to make an alternative release.
I posted the code I redesigned so that you could see and add the options you liked to your release.
I can not put it in separate pieces of code, because this is your code divided into different files, plus my code, plus the code of other developers, a little finished. If interested, download the archive and watch. If you want I can lay out the text files without an archive. https://drive.google.com/drive/folders/0B8OLYcDXEG6rVTdIYmtFMndPdXc?usp=sharing If not interesting, do not use and do not add to the release.

My bad;

Thanks for sharing your implementations. 

I looked over your files quickly.

I like a number of your methods, as well as your email and gotty modules.

The way you have set this up will make a sense to CS, but I think having multiple configuration files like this might be more confusing to many members.

I do think there should be an easy way for members to keep their OC / addresses / settings when updating to a newer 1bash.  A modification of some of your methods might be the opt way to implement this.

I will integrate a lot of this into v0019.   

Before then;

Will you modify the email module to be a single file which: conditionally checks if installed prior, if not installs, if installed executes email based on arguments or files read in?

Basically an email function.

Assume; I would add the email variables to 1bash (other than the email password; this read in from  /home/m1/5_email_password )

Let me know.

[kw1k]

I tried to Change TPccminer to ASccminer for Skein Algo Mining DGB, nothing happend, Mining will not start.

With TPccminer is working well... but AS is faster 20-30%...

Any idea why istn working?

cheers

The linux branch of AS is very far behind the Windows branch; I'm not sure if the linux branch even supports DGB. 



I will look into the link kw1k sent me:

https://github.com/alexis78/ccminer/tree/master

and see if this is different from the version I complied for nvOC.

I compiled the alexis78 git under ubuntu 16.04 VM, this version does support c11. Not sure about DGB algos as I have not tested.

[OverEasy]

I know somewhere buried in 98 pages is the explanation of what Watchdog is but darn if I can find it.

I think there was a link on the OP but don't see it there either.

Can someone please post the link or page number of the info?

What is watchdog watching? How does it work? What the heck is it

See the v0018 demo video:

https://drive.google.com/file/d/0B3ra23kLdr5lSlIxZkJqSm1oWFk/view

Then let me know if I explained it well enough; or if you have more questions.

Yes the video explained it well.
Thanks

[OverEasy]

I am getting errors on V.18
I am SSH ONLY. No monitor. No Guake. No Gnome.

First error is after 1bash tries to stop TeamViewer -


Mon Jul 17 14:26:28 EDT 2017
Action: Removing ...
systemctl stop teamviewerd.service
Failed to stop teamviewerd.service: Unit teamviewerd.service not loaded.
fail

kill '/opt/teamviewer/tv_bin/script/teamviewer' - root (/home/m1/.local/share/teamviewer12 / /home/m1/.local/share/teamviewer12)



Failed to execute operation: No such file or directory
try again, full path...
Failed to execute operation: Invalid argument

Second error is when Maxximus007_AUTO_TEMPERATURE_CONTROL launches -

LAUNCHING:  Maxximus007_AUTO_TEMPERATURE_CONTROL

process in Guake Tab

Failed to connect to Mir: Failed to connect to server socket: No such file or directory
Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help`
       for usage information.

I have to assume Temp Control has failed and is not working.
Can you consider launching temp control and watchdog NOT in Guake if "REMOTE" is chosen but in screens instead? I will try to change bash myself to make this happen.

Thanks

[Nexillus]

Fullzero, IAmNotAJeep or Maxximus007,

Have a question on the watchdog script. I noticed it has triggered twice in the last 24 hours but not sure why as it seems there is not much information within the logs to try to troubleshoot.

The one I am getting is: utilization is too low; reviving did not work so restarting system in 10 seconds. With the restart knocks the miner out for about 20 minutes. I am trying to figure out what is going on and only started having this happen when I went to VER18. I changed the utilization under the script to 80 instead of 90 today to see if that helps.

I appreciate any help or insight into the possible problem, as with the restarts wipes out the other logs.

I have noticed a problem when using both PROFIT_SWITCHING and or WATCHDOG together; I will make changes to the next 1bash to fix it.

It is also possible that the timeouts between checks in the WATCHDOG should be relaxed when mining Ethash COINs.  Or relaxed in general.


If the timing is right; PROFIT_SWITCHING will switch close to the WATCHDOG check and a low utilization will be detected.  

I will make a flag for both IN-SWITCHING and DAG-BUILDING to prevent this in the next 1bash.



What COIN selection are you using or are you using SALFTER_NICEHASH_PROFIT_SWITCHING?

I am not currently using SALFTER_NICEHASH_PROFIT_SWITCHING since Ethash is so close with the profit switching right now.

Not sure what it is but I am rolling back to VER17 on this rig, as it just restarted again and then hard locked up after initial start-up showing a black screen then the flash drive was corrupted and refused to boot.

Just found out it may have been the flash drive the last 48 hours as it just died. LOL write errors all over it. I am testing now with a new flash drive.

What program are you using to image you USB key?

If your using a Windows build computer; I would use hddrawcopy for v0018.

Let me know if this is working with your new USB key.

I am using hddrawcopy on windows 10. It seems to be the flash drive, been over 20 hours now no issues with a new flash drive.

Love new flash drive dies within a week, time to contact newegg!
I am using hddrawcopy  on windows 10, it was the flash drive.

[pixelizedchaos]

So if I want my rigs to be completely headless is there a way to access Teamviewer and enable all those check marks for auto running, and log in through ssh? or is there a better alternative? as it is now, I have a fully closed rig, and I have to rip it apart everytime, I want to update to a new version or if there is an issue with my TV or USB. Thanks!

[hatch789]

Hey Full,

Is there any way you could make the image a tiny bit smaller? I (and I'm sure others) have 16G USB sticks that are already USB 3.0. Unfortunately your image is just a few hundred MB too large to fit onto a 16G stick.

I see there is free space at the end of the image. Any chance you could trim off just 500MB? -If you are able to do that then we can truly use our 16G USB Sticks without hassle.

Also running into an issue with RDP. It won't let me connect. I have tried searching the forum; I know you said it's fully indexed by google... but HOW does one search just this thread in google for something like RDP as an example? If you have an example of that type of search string it would be great.

[Deki1904]

Gparted would be cool. Because I tried to install few things it didn't work for me. Not enough space. I couldn't install Gparted... Would be nice for the next image. Also if you could add solo mining options would be great. That you could connect to one instance server what ever. Maybe for few coins not all at the time. Think it's a lot of work.

Cheers

[hatch789]

Gparted would be cool. Because I tried to install few things it didn't work for me. Not enough space. I couldn't install Gparted... Would be nice for the next image. Also if you could add solo mining options would be great. That you could connect to one instance server what ever. Maybe for few coins not all at the time. Think it's a lot of work.

Cheers

GParted Partition Editor is already included in the build mate! Use the top left button to search and start typing GPAR ...it's there. Trust me. nv0018 build has it at least.

[i0n]

Where is the changelog for v18?

[fullzero]

I tried to Change TPccminer to ASccminer for Skein Algo Mining DGB, nothing happend, Mining will not start.

With TPccminer is working well... but AS is faster 20-30%...

Any idea why istn working?

cheers

The linux branch of AS is very far behind the Windows branch; I'm not sure if the linux branch even supports DGB. 



I will look into the link kw1k sent me:

https://github.com/alexis78/ccminer/tree/master

and see if this is different from the version I complied for nvOC.

I compiled the alexis78 git under ubuntu 16.04 VM, this version does support c11. Not sure about DGB algos as I have not tested.

K good to know.

[fullzero]

I am getting errors on V.18
I am SSH ONLY. No monitor. No Guake. No Gnome.

First error is after 1bash tries to stop TeamViewer -


Mon Jul 17 14:26:28 EDT 2017
Action: Removing ...
systemctl stop teamviewerd.service
Failed to stop teamviewerd.service: Unit teamviewerd.service not loaded.
fail

kill '/opt/teamviewer/tv_bin/script/teamviewer' - root (/home/m1/.local/share/teamviewer12 / /home/m1/.local/share/teamviewer12)



Failed to execute operation: No such file or directory
try again, full path...
Failed to execute operation: Invalid argument

Second error is when Maxximus007_AUTO_TEMPERATURE_CONTROL launches -

LAUNCHING:  Maxximus007_AUTO_TEMPERATURE_CONTROL

process in Guake Tab

Failed to connect to Mir: Failed to connect to server socket: No such file or directory
Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help`
       for usage information.

I have to assume Temp Control has failed and is not working.
Can you consider launching temp control and watchdog NOT in Guake if "REMOTE" is chosen but in screens instead? I will try to change bash myself to make this happen.

Thanks

I will make a conditional execution for the teamviewer block and add the appropriate export cmd in the auto temp block to resolve those errors.

To get auto temp working ( almost positive this is the problem ) add this line:

Code:

EXPORT DISPLAY=:0

I can make each open in a screen when remote is selected.

If you replace the:

Code:

if [ $Maxximus007_AUTO_TEMPERATURE_CONTROL == "YES" ]

block with:

Code:

if [ $Maxximus007_AUTO_TEMPERATURE_CONTROL == "YES" ]
then

cd /home/m1

rm /home/m1/p0
rm /home/m1/p1
rm /home/m1/p2
rm /home/m1/p3
rm /home/m1/p4
rm /home/m1/p5
rm /home/m1/p6
rm /home/m1/p7
rm /home/m1/p8
rm /home/m1/p9
rm /home/m1/p10
rm /home/m1/p11
rm /home/m1/p12
rm /home/m1/p13
rm /home/m1/t0
rm /home/m1/t1
rm /home/m1/t2
rm /home/m1/t3
rm /home/m1/t4
rm /home/m1/t5
rm /home/m1/t6
rm /home/m1/t7
rm /home/m1/t8
rm /home/m1/t9
rm /home/m1/t10
rm /home/m1/t11
rm /home/m1/t12
rm /home/m1/t13
rm /home/m1/fa
rm /home/m1/pa
rm /home/m1/td
rm /home/m1/rpl

sleep 2

echo $INDIVIDUAL_POWERLIMIT_0 > '/home/m1/p0'
echo $INDIVIDUAL_POWERLIMIT_1 > '/home/m1/p1'
echo $INDIVIDUAL_POWERLIMIT_2 > '/home/m1/p2'
echo $INDIVIDUAL_POWERLIMIT_3 > '/home/m1/p3'
echo $INDIVIDUAL_POWERLIMIT_4 > '/home/m1/p4'
echo $INDIVIDUAL_POWERLIMIT_5 > '/home/m1/p5'
echo $INDIVIDUAL_POWERLIMIT_6 > '/home/m1/p6'
echo $INDIVIDUAL_POWERLIMIT_7 > '/home/m1/p7'
echo $INDIVIDUAL_POWERLIMIT_8 > '/home/m1/p8'
echo $INDIVIDUAL_POWERLIMIT_9 > '/home/m1/p9'
echo $INDIVIDUAL_POWERLIMIT_10 > '/home/m1/p10'
echo $INDIVIDUAL_POWERLIMIT_11 > '/home/m1/p11'
echo $INDIVIDUAL_POWERLIMIT_12 > '/home/m1/p12'
echo $INDIVIDUAL_POWERLIMIT_13 > '/home/m1/p13'

echo $TARGET_TEMP_0 > '/home/m1/t0'
echo $TARGET_TEMP_1 > '/home/m1/t1'
echo $TARGET_TEMP_2 > '/home/m1/t2'
echo $TARGET_TEMP_3 > '/home/m1/t3'
echo $TARGET_TEMP_4 > '/home/m1/t4'
echo $TARGET_TEMP_5 > '/home/m1/t5'
echo $TARGET_TEMP_6 > '/home/m1/t6'
echo $TARGET_TEMP_7 > '/home/m1/t7'
echo $TARGET_TEMP_8 > '/home/m1/t8'
echo $TARGET_TEMP_9 > '/home/m1/t9'
echo $TARGET_TEMP_10 > '/home/m1/t10'
echo $TARGET_TEMP_11 > '/home/m1/t11'
echo $TARGET_TEMP_12 > '/home/m1/t12'
echo $TARGET_TEMP_13 > '/home/m1/t13'

echo $__FAN_ADJUST > '/home/m1/fa'
echo $POWER_ADJUST > '/home/m1/pa'
echo $ALLOWED_TEMP_DIFF > '/home/m1/td'
echo $RESTORE_POWER_LIMIT > '/home/m1/rpl'
echo ""
echo "LAUNCHING:  Maxximus007_AUTO_TEMPERATURE_CONTROL "

HCD='/home/m1/Maxximus007_AUTO_TEMPERATURE_CONTROL'
running=$(ps -ef | awk '$NF~"Maxximus007_AUTO_TEMPERATURE_CONTROL" {print $2}')
if [ "$running" == "" ]
then
EXPORT DISPLAY=:0

if [ $LOCALorREMOTE == "LOCAL" ]
then
guake -n $HCD -r Maxximus007_AUTO_TEMPERATURE_CONTROL -e "bash /home/m1/Maxximus007_AUTO_TEMPERATURE_CONTROL"
  echo ""
  echo "process in guake terminal (f12)"
  echo ""
fi

if [ $LOCALorREMOTE == "REMOTE" ]
then
screen -dmS autotemp $HCD
  echo ""
  echo "process in screen autotemp; attach with: screen -r autotemp"
  echo ""
fi

running=""
fi
fi

it should implement this change (but I haven't tested it)

[fullzero]

Fullzero, IAmNotAJeep or Maxximus007,

Have a question on the watchdog script. I noticed it has triggered twice in the last 24 hours but not sure why as it seems there is not much information within the logs to try to troubleshoot.

The one I am getting is: utilization is too low; reviving did not work so restarting system in 10 seconds. With the restart knocks the miner out for about 20 minutes. I am trying to figure out what is going on and only started having this happen when I went to VER18. I changed the utilization under the script to 80 instead of 90 today to see if that helps.

I appreciate any help or insight into the possible problem, as with the restarts wipes out the other logs.

I have noticed a problem when using both PROFIT_SWITCHING and or WATCHDOG together; I will make changes to the next 1bash to fix it.

It is also possible that the timeouts between checks in the WATCHDOG should be relaxed when mining Ethash COINs.  Or relaxed in general.


If the timing is right; PROFIT_SWITCHING will switch close to the WATCHDOG check and a low utilization will be detected.  

I will make a flag for both IN-SWITCHING and DAG-BUILDING to prevent this in the next 1bash.



What COIN selection are you using or are you using SALFTER_NICEHASH_PROFIT_SWITCHING?

I am not currently using SALFTER_NICEHASH_PROFIT_SWITCHING since Ethash is so close with the profit switching right now.

Not sure what it is but I am rolling back to VER17 on this rig, as it just restarted again and then hard locked up after initial start-up showing a black screen then the flash drive was corrupted and refused to boot.

Just found out it may have been the flash drive the last 48 hours as it just died. LOL write errors all over it. I am testing now with a new flash drive.

What program are you using to image you USB key?

If your using a Windows build computer; I would use hddrawcopy for v0018.

Let me know if this is working with your new USB key.

I am using hddrawcopy on windows 10. It seems to be the flash drive, been over 20 hours now no issues with a new flash drive.

Love new flash drive dies within a week, time to contact newegg!
I am using hddrawcopy  on windows 10, it was the flash drive.

Glad it was a hardware issue; an not a major software problem.  The v0018 watchdog reboots a little too aggressively; but with reasonable OC it should still work effectively.

[fullzero]

So if I want my rigs to be completely headless is there a way to access Teamviewer and enable all those check marks for auto running, and log in through ssh? or is there a better alternative? as it is now, I have a fully closed rig, and I have to rip it apart everytime, I want to update to a new version or if there is an issue with my TV or USB. Thanks!

The simplest solution for now is probably:

https://www.amazon.com/dp/B00B2HOKKQ

I don't use teamviewer; so another member can probably answer if it has cmd line interface or not: if it does I can add automation for setup. 

I've been attempting to finish a: monitor / push / update / dashboard app for a while now.  Keep getting sidetracked keeping up with new coins and integrations + there is an rxOC stack now as well.  Although I do think the contributions so far are extremely helpful and have deserved to be bumped to the front of my improvement stack.

[fullzero]

Hey Full,

Is there any way you could make the image a tiny bit smaller? I (and I'm sure others) have 16G USB sticks that are already USB 3.0. Unfortunately your image is just a few hundred MB too large to fit onto a 16G stick.

I see there is free space at the end of the image. Any chance you could trim off just 500MB? -If you are able to do that then we can truly use our 16G USB Sticks without hassle.

Also running into an issue with RDP. It won't let me connect. I have tried searching the forum; I know you said it's fully indexed by google... but HOW does one search just this thread in google for something like RDP as an example? If you have an example of that type of search string it would be great.

I will add using a 15.5gb image to the list.

first enable / configure with the Desktop Sharing GUI on the left on the rig you want to connect to.

then with another ubuntu computer

Try using the Remmina client to RDP.

see:

https://github.com/FreeRDP/Remmina/wiki/Remmina-User%27s-guide

for more info.

[fullzero]

Gparted would be cool. Because I tried to install few things it didn't work for me. Not enough space. I couldn't install Gparted... Would be nice for the next image. Also if you could add solo mining options would be great. That you could connect to one instance server what ever. Maybe for few coins not all at the time. Think it's a lot of work.

Cheers

GParted Partition Editor is already included in the build mate! Use the top left button to search and start typing GPAR ...it's there. Trust me. nv0018 build has it at least.

hatch789 is right; gparted has been installed since ~ v0015.

[fullzero]

Where is the changelog for v18?

its on the OP ( Original Post or post # 1 )

v0018
  re - added SSD / HDD support
  updated driver to: 384.47
  added CLAYMORE 9.7
  added plus integrated SRR support: Thanks xleejohnx  Grin
  added plus integrated IAmNotAJeep_and_Maxximus007_WATCHDOG: Thanks IAmNotAJeep  and  Maxximus007  Grin
  added plus integrated Maxximus007_AUTO_TEMPERATURE_CONTROL: Thanks Maxximus007  Grin
  added plus integrated SALFTER_NICEHASH_PROFIT_SWITCHING: Thanks salfter  Grin
  added AUTO_REBOOT selector
  added GPUPowerMizerMode_Adjust and selector
  added LOCALorREMOTE selector
  added lspci and nvidia-smi output to the top of mining process
  prepended 'nv' to auto workernames
    Zpool_Blake_2S COIN selection
    Zpool_Lyra2v2 COIN selection
    GRS COIN selection
    DMD COIN selection

[fullzero]

Hey guys I am running v0013 and it was running smoothly for 5 weeks.  Now every few days the terminal locks up with this error :

Error message : NVML: cannot get current temperature, error 15

Has anyone here seen the issue before or found a way to resolve?

My rig uses 9x 1070s OC settings are cc100 / mem900 which was stable for 3 weeks.  Recently I power limited to 140 ( I dual mine ) but I seem to get the error whether power limit is on or off.

Since the system freezes I cant remote in to reset it.   I might buy a wifi power point so I can reaet it today but would rather a cure

I scoured the forums and some people reckon turning off WD worked in windows but cant seem to replicate the fix in Linux.

Here was one solution :
"Problem solved : -write "-wd 0" in claymore start.bat -in windows power management high, pci off"

If you only started having this problem after increasing your powerlimit: your risers might not be quite able to handle the higher powerlimit.

What kind of risers are you using, and how are you powering them?


If you look in the system log after a crash:

top left type system, click System Log

it should tell you what likely caused the trash.


If you are using claymore:  I highly recommend updating to v0018 as it has the 9.7 version of claymore which has significantly higher hashrates.

[fullzero]

I got a BIOSTAR TB250-BTC PRO (12x gpu) today Link

I made a 12x 1060 rig with it.

ensure Mining Mode is enabled in the bios LINK  

ensure Max TOLUD is set to 3.5 GB in the bios LINK

NOTE: you must first only connect 6x GPUs, boot, make Bios changes, save and reboot, shutdown, add the other 6x GPUs, attach the USB or SSD and boot





I like the 13x out the box + m2 ssd ready ASRock more; but this is also a good mobo.  

Biostar sadly still can't handle; making mining settings the default.