Building a rock solid multi-gpu linux mining rig with CEntOS 6.0

[Viceroy]

Live long Bitcoin and Centos![/b]

Do what he said, papa and YES YES YES.  ;-)

[RHWN]

Ok, i have a usb flash drive where i have a .iso of CentOS 6.0 LiveCD.
When i boot it up with the usb i get the error after the logo and spinning circle i get this error:
"Root device not found
Sleep forever"

I have booted both ubuntu, win8, magicpartion (?) with this mobo.
Havent changed anything since i installed ubuntu on the machine.

Have a ssd connected to the mobo with sata & no extra usb/ssd/hdds connected.

Dont know what to do... 

[Viceroy]

It is not recognizing your flash drive.  Or it is saying it is bad.  Try booting without the flash drive... does it go to ubuntu?  If so reinstall the ISO on the drive.  Did you use penlinux?

[RHWN]

It is not recognizing your flash drive.  Or it is saying it is bad.  Try booting without the flash drive... does it go to ubuntu?  If so reinstall the ISO on the drive.  Did you use penlinux?

The ssd is erased. Nothing on there anymore.

I used unetbootin. Thougt it woulnt matter, but will try with penlinux!

[Viceroy]

Use the one in the tutorial, that one worked for me. 

[RHWN]

Use the one in the tutorial, that one worked for me.  

Made a bootable usd with the cent os iso nd the program in the guide.
Will try to boot it up now... the usb is also 8gb so should be enuff space...


Edit: Still the same result...

Boot failed, no root device found
Sleep forever

Dont know what to do now...

Edit 2:

Cant get it to work, will just run with win7 

[Viceroy]

sorry you are having trouble boss.  I've used the tool in the guide a bunch of times and have not had any problem.  You could go old-school and burn a cd or DVD. 

you could also go back to ubuntu and just unpack the catalyst drivers as I have recommended.  For your purposes the main difference between ubutnu and centos is the way you update.  ubuntu uses apt-get whereas centos uses yum.  otherwise either will work for basic mining.  centos seems to be the only solution for dual gpu cards, though.

[Crypt_Current]

Use the one in the tutorial, that one worked for me.  

Made a bootable usd with the cent os iso nd the program in the guide.
Will try to boot it up now... the usb is also 8gb so should be enuff space...


Edit: Still the same result...

Boot failed, no root device found
Sleep forever

Dont know what to do now...

Edit 2:

Cant get it to work, will just run with win7 

What brand mobo you using?  If it's Gigabyte, they are infamous for having random sudden problems booting from USB ...

hey Viceroy, +1 for this guide; I burnt my CentOS 6 iso to cd just now and am going to give it a go after next payout 

[Viceroy]

cool, lemme know if you have any trouble.

[spudy12]

Thanks for this guide its really useful!

one thing, I've got to the stage where

####  Install generated package
cd /home/miner/miningpacks/
yum -y install fglrx64_p_i_c-* --nogpgcheck
aticonfig --adapter=all --initial --force
reboot

when i try and run

yum -y install fglrx64_p_i_c-* --nogpgcheck

it no package found?

any ideas what i've done wrong?

[Viceroy]

> when i try and run   yum -y install fglrx64_p_i_c-* --nogpgcheck
>it no package found?
>any ideas what i've done wrong?

The file fglrx64 is a file generated when you install the driver.  This is a machine specific file that was built just for your machine.  It appears to be missing.  One possible cause of that is that your machine is not running at 64 bits.  If you have a 32 bit machine (pretty common) your file *might* be named fglrx32 or fglrx.

After "making" the machine specific driver do an "ls" command to see what is in the directory.  

I'm guessing:
1) you don't have a 64 bit machine so it didn't build correctly, maybe not at all
2) you are in the wrong directory, the files is somewhere else
3) you built a different version of the driver that has some other name


useful commands at this time are:
ls

The "ls" command "lists" what is in the current directory.

you may also want to run this as root:
updatedb

then run:
locate fglrx64

On my machine the output looks like this:

Code:

[root@localhost ~]# locate fglrx64
/home/miner/miningpacks/fglrx64_p_i_c-8.841-1.x86_64.rpm
/root/rpmbuild/RPMS/x86_64/fglrx64_p_i_c-8.841-1.x86_64.rpm
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/checksum_data
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/checksum_type
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/command_line
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/from_repo
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/from_repo_revision
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/from_repo_timestamp
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/reason
/var/lib/yum/yumdb/f/8db4396703e2b87f70da7cd1caa8aeb876ca07cf-fglrx64_p_i_c-8.841-1-x86_64/releasever
[root@localhost ~]#

clearly I have "located" the file I need in the correct directory here:
/home/miner/miningpacks/fglrx64_p_i_c-8.841-1.x86_64.rpm

[spudy12]

okay thanks for that! I'm defiantly running the 64 bit version, I'll try and reinstall the driver again then try the locate command.
thanks!

[Viceroy]

see if you have any problems during the driver creation phase, look for error messages...

[cr1776]

Hi,
I had a couple of comments on the guide, if you are looking for feedback.  I don't think it is my error,
but if so, my apologies!  (I was using the instructions on page 1, note later in the thread, hope that was the current one).

This has been helpful, although I am not yet successful in running it yet - I went from Ubuntu 10.10 to upgrade from getwork thinking it would not be difficult as it was not when I set up 10.10 several years ago, but this has been frustrating.  (With Ubunutu I was not having success getting the drivers etc set, so thought I'd try CentOS 6.0 ).

I'm hoping I'll have some success shortly, but am stuck right now.


=====================================================================
1.  Looks like a typo:

At one point you have "3point0" and then later reference "5point0".  Just FYI:
cp /home/miner/miningpacks/ADL-SDK-3point0-temp/include/*.h /home/miner/cgminer/ADL_SDK
rm -rf /home/miner/ADL-SDK-3point0-temp

cd /home/miner/cgminer/ADL_SDK
cp /home/miner/miningpacks/ADL-SDK-5point0-temp/include/*.h .
rm -rf /home/miner/miningpacks/ADL-SDK-5point0-temp/



=====================================================================
2.
Also in one spot, you need to add a mkdir to make:
mkdir /home/miner/cgminer/
mkdir /home/miner/cgminer/ADL_SDK

before doing this:
cp /home/miner/miningpacks/ADL-SDK-3point0-temp/include/*.h /home/miner/cgminer/ADL_SDK
rm -rf /home/miner/ADL-SDK-3point0-temp



BUT later, if you do that git will complain that the cgminer directory exists.



=====================================================================
3. Regarding the installation path for the ATI Catalyst driver install in Step 6, the default is /
Is that what you had used for the installation path?

Just notes for some others:
You want "Package Generation " and "RedHat Packages" and then "RedHat/RHEL6_64a" (you need to hit SPACE to select it in the terminal (not running it in the GUI, but ssh'd in, then return).

Then you want "Generate Distribution Specific Driver Package" (not Install Driver 8.841 on X.Org 6.9 or later 64-bit".)

         


=====================================================================
4. What I am seeing right now, I think it is the kernel issue you had mentioned about non-matching
kernels maybe, but I am attempting to figure out how to fix it:

#aticonfig --list-adapters
* 0. 13:00.0 AMD Radeon HD 6990
  1. 12:00.0 AMD Radeon HD 6990
  2. 0b:00.0 AMD Radeon HD 6990
  3. 0a:00.0 AMD Radeon HD 6990
  4. 07:00.0 AMD Radeon HD 6990
  5. 06:00.0 AMD Radeon HD 6990
 
 
$yum list installed | grep kernel-*
dracut-kernel.noarch                004-33.2.el6_0                     @updates
kernel.x86_64                       2.6.32-71.el6                      installed
kernel.x86_64                       2.6.32-71.29.1.el6                 @updates
kernel-devel.x86_64                 2.6.32-71.29.1.el6                 @updates
kernel-firmware.noarch              2.6.32-71.29.1.el6                 @updates
kernel-headers.x86_64               2.6.32-71.29.1.el6                 @updates

$ uname -r
2.6.32-71.29.1.el6.x86_64




$./cgminer -n
No protocol specified
 [2013-05-28 11:05:26] CL Platform 0 vendor: Advanced Micro Devices, Inc.                   
 [2013-05-28 11:05:26] CL Platform 0 name: AMD Accelerated Parallel Processing                   
 [2013-05-28 11:05:26] CL Platform 0 version: OpenCL 1.1 AMD-APP-SDK-v2.4 (595.10)                   
 [2013-05-28 11:05:26] Error -1: Getting Device IDs (num)                   
 [2013-05-28 11:05:26] clDevicesNum returned error, no GPUs usable                   
 [2013-05-28 11:05:26] 0 GPU devices max detected 
 
 
 

[cr1776]

I do see this too:

$lspci -nn | grep VGA
07:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:671d]
0b:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:671d]
13:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:671d]


$ps aux | grep X
root      3333  0.2 10.1 182620 188628 tty1    Ss+  11:42   0:06 /usr/bin/Xorg :0 -nr -verbose -auth /var/run/gdm/auth-for-gdm-BjwGpq/database -nolisten tcp vt1


sudo DISPLAY=:0 fglrxinfo
display: :0.0  screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6990
OpenGL version string: 4.1.10666 Compatibility Profile Context



display: :0.0  screen: 1
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6990
OpenGL version string: 4.1.10666 Compatibility Profile Context



display: :0.0  screen: 2
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6990
OpenGL version string: 4.1.10666 Compatibility Profile Context



display: :0.0  screen: 3
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6990
OpenGL version string: 4.1.10666 Compatibility Profile Context



display: :0.0  screen: 4
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6990
OpenGL version string: 4.1.10666 Compatibility Profile Context



display: :0.0  screen: 5
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6990
OpenGL version string: 4.1.10666 Compatibility Profile Context

[Viceroy]

Beautiful and I will incorporate your changes asap. 

Can you see glxgears?

$ glxgears
6564 frames in 5.0 seconds = 1312.776 FPS
39798 frames in 5.0 seconds = 7959.491 FPS

Are you at the head?  Or remote?  What's your setup?

let me know what this returns:

env DISPLAY=:0.0 aticonfig --pplib-cmd "set fanspeed 0 100"

or

env DISPLAY=:0.1 aticonfig --pplib-cmd "set fanspeed 0 100"

[cr1776]

Hi,
I am ssh'ing in right now, but can go to it if needed.  It has been running nicely for about 2 years now until I decided to "upgrade" on Friday.  3 AMD Radeon HD 6990s.  2 GB RAM.  500GB HDD.

Thanks for the reply, regarding your questions:

1.
$ glxgears
49458 frames in 5.0 seconds = 9891.541 FPS
49647 frames in 5.0 seconds = 9929.385 FPS
49683 frames in 5.0 seconds = 9936.522 FPS
...


2.
env DISPLAY=:0.0 aticonfig --pplib-cmd "set fanspeed 0 100"
PPLIB command execution is Successful!

3.
env DISPLAY=:0.1 aticonfig --pplib-cmd "set fanspeed 0 100"
PPLIB command execution is Successful!

Beautiful and I will incorporate your changes asap. 

Can you see glxgears?

$ glxgears
6564 frames in 5.0 seconds = 1312.776 FPS
39798 frames in 5.0 seconds = 7959.491 FPS

Are you at the head?  Or remote?  What's your setup?

let me know what this returns:

env DISPLAY=:0.0 aticonfig --pplib-cmd "set fanspeed 0 100"

or

env DISPLAY=:0.1 aticonfig --pplib-cmd "set fanspeed 0 100"

[Viceroy]

and
.cgminer -n

just fails, huh?

That's what I was getting before I wrote the script.  Did you absolutely make sure you cut off the yum upgrades?  Because yum is not friendly for this, if you did not modify the yum conf file before running a yum update/upgrade that will break it.

[cr1776]

Yeah, just bizarre. I believe I did, but did it twice.  I will re-try one more time.  ;-)


$ ./cgminer -n
No protocol specified
 [2013-05-28 11:05:26] CL Platform 0 vendor: Advanced Micro Devices, Inc.                   
 [2013-05-28 11:05:26] CL Platform 0 name: AMD Accelerated Parallel Processing                   
 [2013-05-28 11:05:26] CL Platform 0 version: OpenCL 1.1 AMD-APP-SDK-v2.4 (595.10)                   
 [2013-05-28 11:05:26] Error -1: Getting Device IDs (num)                   
 [2013-05-28 11:05:26] clDevicesNum returned error, no GPUs usable                   
 [2013-05-28 11:05:26] 0 GPU devices max detected 
 

and
.cgminer -n

just fails, huh?

That's what I was getting before I wrote the script.  Did you absolutely make sure you cut off the yum upgrades?  Because yum is not friendly for this, if you did not modify the yum conf file before running a yum update/upgrade that will break it.

[Viceroy]

I do not know if this is important but  my results differ from yours very slightly:

Code:

 yum list installed | grep kernel-*
dracut-kernel.noarch                004-32.el6                         installed
kernel.x86_64                       2.6.32-71.el6                      installed
kernel-devel.x86_64                 2.6.32-71.el6                      @base
kernel-devel.x86_64                 2.6.32-71.29.1.el6                 @updates
kernel-firmware.noarch              2.6.32-71.el6                      installed
kernel-headers.x86_64               2.6.32-71.29.1.el6                 @updates