Topic: [OS] nvOC easy-to-use Linux Nvidia Mining - page 42. (Read 418253 times)

as a first thank you for the many work you do here! !!!

I love nvOC it makes life really easier :-)
I use it now on all rig's

Are there any plans when NXS Nexuscoin will be intigrated with algo SHA3?

linux is still not easy for me
compiling kills me

but I could offer help with testing gpu minern for sha3 or others or new stuff.
I have access to various nv gpus 1070 1070ti 1080 + ti

Last but not least, my English is not as good as yours but I give it my best

Try this and get back with the output:

lspci -nnk | grep 0200 -A3

There seems to be a issue with the hardware if the lspci and lshw are returning naught...

Being a Gigabyte mobo, known for their hairy implementations of the Realtek to ICH bridging, it might help also to update your BIOS:

https://www.gigabyte.com/MicroSite/462/images/BIOS-Z270P-D3.zip

NOTE: Please check the HW revision of your mobo before trying to update. This is a "mining" optimised BIOS

I've seen  something similar a while back with a rev. 1 board, similar to yours, that was running out of lanes for peripherals and the loading of the PHY drivers for ethernet and usb were loaded erratically at times.

The Community Release edition was based and built upon nvoc 19-1.4 which is built on top of Ubuntu 16.04. If I understand correctly the vBASIC version is also built on top of Ubuntu 16.04. This should not be specific problem tied to the Community Release and if in fact it is related to base Ubuntu 16.04 than it will affect both, the Community Release and vBASIC release. It seems like his motherboard is defective.

You pretty much answered your question. It seems something is wrong with your motherboard. Have you had the same problem before flashing the new bios? Try downgrading the bios, the latest one might be buggy. If it doesn't help then return or RMA the board while you still have warranty.

Initially trying to mine SMART.  When you say vBasic, is that v19?  --DISREGARD fullzero, I saw your post a few back from this and know what you are referring to.  I truly don't think it is the OS (or, if it is, the "trying a different OS ship" has already sailed as I've royally screwed enough stuff up along the way...a total fresh start from scratch is likely the only fix), but I will certainly give the vBASIC a try.  Thank you. --

I have not tried installing Windows.  But even if I interrupt the boot sequence and look at the System Information, the MAC address for the LAN controller (which is onboard to the motherboard) is blank.  It's like the motherboard doesn't see it.  And if I try to plug a USB WiFi adapter in after boot-up, it's dead too.  No activity at all.  Yet if I plug a USB data stick in to the same USB port, it opens up and lets me browse the stick.  Regardless of the OS, wouldn't the fact that the BIOS interrupt and review of system info not showing an active LAN controller indicate it is an issue prior to the OS?

What are you trying to mine?  If you are planning on mining an Ethash coin; If you are having trouble with the community edition: I recommend trying vBasic as that motherboard should be supported by it.  

It's very unusual for network card not to be recognized by nvoc. Not impossible, but unusual.
Since you have extensive Windows experience, have you tried installing Windows on that motherboard to confirm it is not a hardware failure?
If your network card works well under Windows we can rule out defective hardware and try to troubleshoot further.

You're right, as I said it depends on how much a noob he feels, but if he really need dual-booting I had pleasure in giving him some hints. As you said, I think the quickest way to do it is the first option, nvOC first then windows. About the RAID, I remember that all my old intel mobo's which have onboard RAID controllers available, all worked the same way and would allow you to create mirrored/striped RAID volumes across arbitrary disks portions shaping what Intel called a "Matrix", even if you choose to do a full raid0 or full raid1 across two disks you will see a virtual disk (a "vector" of the matrix) in boot menu and os instead of the two real one. Maybe it wouldn't allow to create "scalar" volumes, i.e. non-RAID volumes from a single disk.

Putting all my responses in RED and in one reply (working from nvOC in LOCAL mode so not able to copy-and-paste results, so I'm slow...)


When I run that command, it flashes "PCI (sysfs)" real quickly & disappears and then flashes "USB" real quickly & disappears and then returns to the command prompt of m1@m1-sesktop:~$


When I run that command same outcome as command above: it flashes "PCI (sysfs)" real quickly & disappears and then flashes "USB" real quickly & disappears and then returns to the command prompt of m1@m1-sesktop:~$


This does nothing other than return to the command prompt.


Returns these three lines:
     # interfaces(5) file used by ifup(8) and ifdown(8)
     auto lo
     iface lo inet loopback



I get the following message:  insserv: warning: script 'gpu' missing LSB tags and overrides

Then I used:  sudo service network-manager restart
That showed a graphical message that said I was now disconnected and off-line (which is interesting since I wasn't on-line to begin with

Yes, I'm using Guake! terminal.

And to your question about running lspci -nnk | grep -iA2 eth, that returns nothing; it just returns to a command prompt and sudo cat /etc/network/interfaces yields:

# interfaces(5) file used by ifup(8) and ifdown(8)
auto lo
iface lo inet loopback

Will head to Discord

R u running the command using the guake terminal ?

Come over to discord and lets work it out over there.

When I run that command, it flashes "PCI (sysfs)" real quickly & disappears and then flashes "USB" real quickly & disappears and then returns to the command prompt of m1@m1-sesktop:~$

In order to be able to help you a bit, try to run the following command:

 sudo lshw -C network

then

sudo lshw -C network | grep logical

and paste here the result of both commands.

For some even more debug options, the result of the following commands will give us a hint on the potential driver/modules issues:

sudo lspci -nnk | grep -iA2 eth

sudo cat /etc/network/interfaces



You can always try to re-initialize the network-manager using this command:

sudo dpkg-reconfigure network-manager

Am a newbie to mining and setting up my first rig.  Have an extensive Windows background.  Have zero Linux background (except for the immersion course I'm in the middle of trying to get nvOC v2.0 set up and running ).

I got nvOC up and running but my issue is, somewhere along the way, the LAN controller ceased being recognized.

My setup:

• Attempting a 6-card build with all 1080ti's, largely following Vosk's "How to Build a 6x 1080 TI Mining Rig with ATX & Server PSU's" (https://www.youtube.com/watch?v=2MxNVySpbcw)
• Motherboard = Gigabyte Z270P-D3
• PSU (to power risers) = EVGA Supernova 650 P2
• PSU (to power cards) = Bitmain APW3++
• Harddrive = Lexar 32G USB 3.0 memory stick

I'm initially working to get one card set up and running.  It is plugged into a riser that is connected to the primary PCI slot on the mobo, both of which (riser and card) are powered by the EVGA PSU (once I get one card running I'll work up to all six cards and then power the risers solely with the EVGA and the cards with the Bitmain).  I have the most current BIOS for the mobo installed, 1bash is configured to mine SMART and I can boot & have the miner software functioning.  But my current issue is that the LAN controller is not visible to the OS.

I've had fits and starts thru this process, re-imaging the USB probably 4 or 5 times, for various things I hadn't done correctly (e.g., hadn't updated the BIOS before booting, had plugged my monitor into the mobo HDMI instead of the cards's HDMI and dealt with Xorg issues, etc.).  I know that at one point, the LAN controller was working (green light flashing on controller, which BTW, is a Realtek® GbE LAN chip (10/100/1000 Mbit))-- I'm just not sure when it quit working to be able to exactly determine what I might have changed that caused the issue.

Diagnostics & Steps I've Taken To Address

• Re-imaged nvOC
• Reset CMOS (including pulling mobo battery for about 30 minutes)
• ifconfig -a only shows a "lo" entry, no WLAN
• sudo lshw -c network flashes "PCI (sysfs)" real quickly and then flashes "USB" real quickly and then returns to the command prompt
• And yes, I have tried plugging in an ethernet cable that was live and working in another computer sitting right next to my rig

I guess the issue could be a fried controller, but this is a brand new mobo.  The only other thing I did differently at one point was change the mouse I was using from a USB mouse to an older wireless mouse where the wireless receiver plugs into the light green mouse port on the back of the mobo instead of being a USB mouse.  But I've re-imaged, reset CMOS, etc. since then and even went without using a mouse on a couple of attempts to see if that had caused some issue.

Anyway, I would greatly appreciate any assistance and guidance the Community might have to offer.

Thank you!

Sorry, my bad. No need to escape + in this case. Any of these 4 examples should work:

Didn´t work because it puts the "\" into the address like this
and that fails because the pool is not expecting to see the "\" in there. It only expects to see the "+" expression in between the etnaddress and the workername.

You have to escape + sign with \
Try:

It can be done, however it takes a lot of steps and advanced user to do it. This is beyond the simplicity of nvoc's plug-n-play structure and general support from the community. We can't support custom solutions.

If it's just to evaluate nvoc the easiest way is to burn it on USB stick and try if it performs well for you. If it does, consider installing it on dedicated SSD. Otherwise, what's the point of dual booting on dedicated mining rig that will mine 24/7?

If you still want to dual boot (let's say you have some miner that works on windows only), the easiest solution is to install nvOC first, then create new partition and install Windows on it and update grub bootloader to list windows as well. Installing nvoc as secondary OS on SSD that already has Windows it's more complicated.

The other options listed, like cloning nvoc from git on freshly installed Ubuntu won't work properly as there are other manually tweaked settings to Ubuntu (e.g. cuda environment variable, autostart, auto-update settings...) and extra packages installed (cuda, nvidia drivers, compilers...) that are not included on git.

I am not aware of raid level that lets you split hdd to see it as two separate hard drives. Usually raid is used for mirroring and redundancy and not all motherboards have raid controller built in.

Can you please give an example, illustrating the above statement, sorry i didn't quite get that to be honest.

So, this as well just a general explanation (not just for you but for any new miners to understand) of how actually 1bash, 0miner or 3main works.

We add all the coin details and they are stored/attached to variables.

We call those variables in 3main or 0miner and shape them up according to how pool expects.

We follow the same procedure for every single coin.

Feel free to ask any further questions you have.

[/quote]

Well the etn pools use the following syntax:

where xxxxx represents the fixed diff

where yyyyy represents the workername and

where both are implemented the fixed diff represented by xxxxx and the workername represented by yyyyy so that is what i meant by using a +.

To add the workername to an address but not setting the miner to a fixed diff one could use the "+" sign between the etn address and the workername. I tried to implement this into my 0main but failed because of my limited linux knowledge. I think that the + sign has a meaning when used with variables in a bash script but i could not get it to work.

ADDR=$((ETN_ADDRESS+$ETN_WORKER)) gave me a value too great for base error (error token = the expression of my workername)

Then again it is not that important i just liked to use the telegram script in nvOC and to differentiate between future machines i gave them significant names. By now we know that my name starting with the 19 used with the stock 0main in the community release led to my problem with the fixed diff so i do not know how to implement this correctly into the 0main script but again i think there is more important stuff for you guys than this.