nvidia nv issues - notes#2

az group create -l eastus -n azurebigcompute && az group deployment create -g azurebigcompute -n azurebigcompute --template-uri https://raw.githubusercontent.com/Azure/azure-bigcompute/master/azuredeploy.json --parameters "{\"singleOrCluster\":{\"value\":\"single\"},\"DnsLabelPrefix\":{\"value\":\"ahmedmosgpunv\"},\"AdminUserName\":{\"value\":\"ahmedmos\"},\"SshPublicKey\":{\"value\":\"ssh-rsa AAAAB3NzaC1yc2EAAAABJQAAAQEAuYLenQ/XeHxwShcDQVSxuxCW7irgqcAbU63kiDCw+hp7SZr5eo7AYFtqd4+cor8jek7/Qxw8Dx14pjxC5PrON68ZlNnf06arMOeE7Pwdr9uVPIWLFiOgnZhtHw72SWt5NeRjJv8B8CLsZgewxtFTM+gtwZcTLfXA//HpqwvduVuBsLfQzaHkAl+KP/NM+LMWdobk3eaJWeHTFbKT4fORIzWbW7cWHb1pvZU3qhPGcEwWnLujhsC14pkGbclw2zCi0bu5jM5/4whjHM6reGbejZZgZdxa9L+x/R4kDHxf5TnONP841XbdkxYmxLAyqMMhDZLn6fpsp0lFMuVT95aJbw== ahmedmos-key-nopass\"},\"ImagePublisher\":{\"value\":\"openlogic\"},\"ImageOffer\":{\"value\":\"CentOS\"},\"ImageSku\":{\"value\":\"7.3\"},\"HeadandWorkerNodeSize\":{\"value\":\"Standard_NV24\"},\"WorkerNodeCount\":{\"value\": 0},\"GpuHpcUserName\":{\"value\":\"gpuuser\"},\"NumDataDisks\":{\"value\":\"32\"}}" --debug

AzCopy /Source:https://tdcm16sg112leo8193ls102.blob.core.windows.net/tdcm16sg112leo8193ls102/ /Dest:C:\Users\ahmedmos\Downloads\AzureBigComputeStuff\ /SourceKey:M4TbEmWQOw5EOV4JI3cOlei9d3LfNwW7TSe0pTYK1AO4Lns/5mBdnbceEzyAigWDrmvNiHPk7oywSeOpBPr+Fg== /S

storage account & container tdcm16sg112leo8193ls102
AzCopy /Source:C:\Users\ahmedmos\Downloads\AzureBigComputeStuff\ /Dest:https://tdcm16sg112leo8193ls102.blob.core.windows.net/tdcm16sg112leo8193ls102/ /DestKey:HR1J/klZfC69wq47+E6RR2CzaOPoGeeDGzYBgfm4YYimnjUvztS61pQPDUWld9Rl6JvxWg6ItwXzCTyur29/Jw== /S

run after provisioning (everything runs in root):

yum install -y kernel kernel-devel kernel-debug kernel-headers
run the code in lines https://github.com/Azure/azure-bigcompute/blob/master/azuredeploy.sh#L833-L856

reboot BEFOR DRIVER INSTALL

wget https://tdcm16sg112leo8193ls102.blob.core.windows.net/tdcm16sg112leo8193ls102/azurenvidia42/NVIDIA-Linux-x86_64-367.103-grid.run

check kernel version matches
[root@ahmedmos ~]# uname -a
Linux ahmedmos 3.10.0-514.16.1.el7.x86_64 #1 SMP Wed Apr 12 15:04:24 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
[root@ahmedmos ~]# rpm -q kernel
kernel-3.10.0-514.2.2.el7.x86_64
kernel-3.10.0-514.16.1.el7.x86_64
[root@ahmedmos ~]# rpm -q kernel-devel
kernel-devel-3.10.0-514.16.1.el7.x86_64
[root@ahmedmos ~]# rpm -q kernel-debug
kernel-debug-3.10.0-514.16.1.el7.x86_64
[root@ahmedmos ~]# rpm -q kernel-headers
kernel-headers-3.10.0-514.16.1.el7.x86_64

yum erase kernel-3.10.0-514.2.2.el7.x86_64

chmod +x NVIDIA-Linux-x86_64-367.92-grid.run
dracut --force


#change the graphical target:
init 3

rpm -Uvh ftp://rpmfind.net/linux/fedora/linux/development/rawhide/Everything/x86_64/os/Packages/d/dkms-2.3-4.20170313git974d838.fc27.noarch.rpm
yum install -y dkms

./NVIDIA-Linux-x86_64-367.92-grid.run --silent --dkms --install-libglvnd

[root@ahmedmos ~]# nvidia-smi
Wed Apr 26 11:10:05 2017
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 367.92                 Driver Version: 367.92                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla M60           Off  | 99F3:00:00.0     Off |                  Off |
| N/A   39C    P0    39W / 150W |      0MiB /  8123MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla M60           On   | A33A:00:00.0     Off |                  Off |
| N/A   37C    P0    39W / 150W |      0MiB /  8123MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla M60           On   | A689:00:00.0     Off |                  Off |
| N/A   32C    P0    39W / 150W |      0MiB /  8123MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   3  Tesla M60           On   | A9C4:00:00.0     Off |                  Off |
| N/A   41C    P0    40W / 150W |      0MiB /  8123MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

yum install -y tigervnc-server

#check the gridd service 
switch back to ahmedmos
sudo systemctl status nvidia-gridd
sudo systemctl start nvidia-gridd

#update the vnc config as tab new9 

vncserver
enter a password (Nothing_00)

#download vnc client and enter the machine address with :1 then enter the password

you will see GUI and go to nvidia driver, you will see it not running

sudo killall Xvnc
sudo nvidia-xconfig

sudo su -
vi /etc/X11/xorg.conf



This part goes into vnc ~/.vnc/xstartup configuration
--------------

#!/bin/sh

# Add the following line to ensure you always have an xterm available.

( while true ; do xterm ; done ) &

# Uncomment the following two lines for normal desktop:

unset SESSION_MANAGER

exec /etc/X11/xinit/xinitrc

[ -x /etc/vnc/xstartup ] && exec /etc/vnc/xstartup

[ -r $HOME/.Xresources ] && xrdb $HOME/.Xresources

xsetroot -solid grey

vncconfig -iconic &

xterm -geometry 80×24+10+10 -ls -title “$VNCDESKTOP Desktop” &

#twm &

startkde &


------------
this will go into /etc/X11/xorg.conf (don't add it now because you don't know if it works, check with MS & nvidia)

Section "DRI"

	Mode 0666

EndSection

Section "Device"

    Identifier     "Device0"

    Driver         "nvidia"

    VendorName     "NVIDIA Corporation"

    BoardName      "GRID M60-4Q"

    BusID          "PCI:2:0:0"

EndSection
---------------------

echo ‘IgnoreSP=TRUE’ | sudo tee -a /etc/nvidia/gridd.conf
wget -O /etc/yum.repos.d/pcoip.repo https://downloads.teradici.com/rhel/pcoip-beta.repo
wget https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm
rpm -i epel-release-latest-7.noarch.rpm
yum update
yum install -y pcoip-agent-graphics
systemctl status pcoip
systemctl enable pcoip

pcoip-register-host --registration-code=ECAUYB969PWS@DE86-369B-88A7-4940


now download pcoip client from aka.ms/accessgpu and try connecting:

question to teradici:
	Can you provide me with steps/documentation to connect over pcoip without password to my linux NV24 VM

stop pcoip 
exit
on ahmedmos user, start vncserver

download opendtect and chmod+x and install it via GUI

# add vnc config to user gpuuser
#give access to the 32TB  
sudo su -
cd /
chown -R gpuuser:gpuuser /data

sudo su gpuuser
mkdir DTECT_DATA

install OPENDTECT