Creating an Amazon Machine Image for Flame Family Products (CentOS 7.6)

Important:

This topic explains how to create an instance from scratch, using CentOS 7.6. With Flame Family 2023.1 Update, creating an instance with Rocky Linux 8.5 is simpler, and is described here.

This topic explains how to create an Amazon Machine Image (AMI) to run an instance loaded with an Autodesk Flame Family product: Flame, Flare, or Flame Assist.

You need to create an Open Virtualization Archive (OVA) with only a basic Operating System (OS) installation. You then use this OVA to create the AMI that holds the software to run on an AWS instance. These instructions are only a baseline for you to adapt to your configuration.

Prerequisites

To follow the instructions, you must have:

Create an AMI

To create an AMI compatible with an Autodesk Flame Family cloud deployment, you:

  1. Create an OVA file (a virtual machine) of the base OS with VirtualBox.
  2. Convert the OVA file to a base AMI through AWS services.
  3. Configure the base AMI by installing application-specific packages like DKU, and Autodesk Flame.
  4. Create the AMI you will use to deploy a Flame Family product.

Creating the Base OS OVA

When creating an AMI from scratch, the first step is to create a basic OS-only AMI from the Autodesk CentOS 7.6 iso. You then install the application-specific packages (DKU, Flame) on the new image.

You need the Autodesk CentOS 7.6 ISO as the source OS, available from your Autodesk Account.

  1. Copy the source iso to your local machine.

  2. In VirtualBox Create a new OVA with the following settings.

    1. Name: adsk_flame_workstation_centos_76
    2. Type: Linux
    3. Version: Red Hat (64-bit)

  3. Memory Size: Keep as-is.

  4. Virtual hard disk:

    1. VDI
    2. Dynamically allocated
    3. Size: 50 GB

  5. Storage:

    • Controller: IDE - Optical drive

  6. ISO: Autodesk CentOS 7.6

Once the VM is created, finalize the VM settings:

  1. Start the VM.

  2. Select Autodesk Flame Workstation (manual partitioning).

  3. Partition the disk into /boot (1G) and / (the rest of the disk). Do NOT create a swap partition.

  4. Perform the rest of the OS installation normally.

  5. When the user creation screen appears, create a user named flameadmin:

    • Full Name: Flame Administrator
    • User name: flameadmin
    • Check Make this user administrator
    • Uncheck Require a password to use this account

  6. Once the installation is complete, shut down the VM (File > Close).

  7. In Settings > Storage, delete the IDE Controller but keep the SATA Controller.

  8. Start the VM.

  9. Log in with the user flameadmin.

  10. Disable the root password:

    sudo passwd -l root

  11. Check if your network interface is enabled:

    ip -br addr show | grep UP

    If an IP address appears next to UP, your network interface is enabled. If there is no IP address run the following command, replacing with the first item from the previous command, like enp0s3:

    sudo ifup <interface>

  12. Install cloud support packages and configure keypair login:

    sudo yum install cloud-init cloud-utils-growpart -y
sudo sed -i 's/name: centos/name: flameadmin/' /etc/cloud/cloud.cfg
sudo sed -i 's/gecos: Cloud User/gecos: Flame Administrator/' /etc/cloud/cloud.cfg
echo "preserve_hostname: true" | sudo tee --append /etc/cloud/cloud.cfg
sudo sed -i 's/^PasswordAuthentication yes/PasswordAuthentication no/' /etc/ssh/sshd_config
sudo sed -i 's/^ChallengeResponseAuthentication yes/ChallengeResponseAuthentication no/' /etc/ssh/sshd_config

  13. Shut down the VM.

  14. Export the VM. Make sure to disable Write Manifest File.

Converting the OVA file to an AMI

You now import your OVA in AWS and convert it to the AMI format. Follow the instructions, but for more information, see AWS detailed documentation.

  1. Upload the OVA file you created to an AWS S3 Bucket.

  2. On your local machine, create a text file with the following content, and save it as vmimport-trust-policy.json:

    {
 "Version": "2012-10-17",
 "Statement": [
     {
         "Effect": "Allow",
         "Principal": { "Service": "vmie.amazonaws.com" },
         "Action": "sts:AssumeRole",
         "Condition": {
             "StringEquals": {
             "sts:Externalid": "vmimport"
             }
         }
     }
  ]
}

  3. On your local machine, create a text file with the following content, and save it as vmimport-role-policy.json. Make sure to replace <BUCKET_NAME> on lines 16 and 17 with the name of the S3 bucket used in step 1.

    {
 "Version": "2012-10-17",
 "Statement": [
     {
         "Effect": "Allow",
         "Action": [
             "s3:CreateBucket",
             "s3:DeleteBucket",
             "s3:DeleteObject",
             "s3:GetBucketLocation",
             "s3:GetObject",
             "s3:ListBucket",
             "s3:PutObject"
         ],
         "Resource": [
             "arn:aws:s3:::<BUCKET_NAME>",
             "arn:aws:s3:::<BUCKET_NAME>/*"
         ]
     },
     {
         "Effect": "Allow",
         "Action": [
             "s3:ListAllMyBuckets",
             "iam:CreateRole",
             "iam:PutRolePolicy",
             "sts:AssumeRole",
             "ec2:CancelConversionTask",
             "ec2:CancelExportTask",
             "ec2:CreateImage",
             "ec2:CreateInstanceExportTask",
             "ec2:CreateTags",
             "ec2:DeleteTags",
             "ec2:ImportInstance",
             "ec2:ImportVolume",
             "ec2:StartInstances",
             "ec2:StopInstances",
             "ec2:TerminateInstances",
             "ec2:ImportImage",
             "ec2:ImportSnapshot",
             "ec2:CancelImportTask",
             "ec2:ModifySnapshotAttribute",
             "ec2:CopySnapshot",
             "ec2:RegisterImage",
             "ec2:Describe*"
         ],
         "Resource": "*"
     }
 ]
}

  4. On your local machine, create a text file with the following content and save it as containers.json. Make sure to edit these placeholders:

    • <DESCRIPTION> on line 3 with a description of the AMI.

    • <BUCKET_NAME> on line 6 with your S3 bucket name.

    • <OVA_FILE_NAME> on line 7 with the file name of the OVA file you created.

    [
 {
     "Description": "<DESCRIPTION>",
     "Format": "ova",
     "UserBucket": {
         "S3Bucket": "<BUCKET_NAME>",
         "S3Key": "<OVA_FILE_NAME>"
     }
 }
]

  5. Create the required role and policy. On your local machine, run the following commands:

    aws iam create-role --role-name vmimport --assume-role-policy-document file://vmimport-trust-policy.json
aws iam put-role-policy --role-name vmimport --policy-name vmimport --policy-document file://vmimport-role-policy.json

  6. Give the vmimport role the necessary permissions. In the AWS Web Console, do the following:

    1. In IAM Management Console > Roles > search for vmimport.

    2. Click vmimport.

    3. Under Permissions, click Add Permissions > Attach Policies.

    4. attach the AmazonEC2FullAccess and AmazonS3FullAccess policies

  7. Import the VM. On your local machine, enter the following command in a shell. Make sure to edit these placeholders:

    • <YOUR_AWS_REGION> with your AWS region identifier, such as us-east-1.

    • <YOUR_DESCRIPTION> with a description of the image.

    aws ec2 import-image --region <YOUR_AWS_REGION> --description "<YOUR_DESCRIPTION>" --disk-containers file://containers.json --role-name "vmimport"

  8. The import process can take a while; you can monitor its progress with: 

    aws ec2 describe-import-image-tasks

Once the process is "completed", the new AMI is available under EC2 > AMIs.

Configure the base AMI for Autodesk Flame Family Products

To create a Flame AMI, we will start from the base AMI created in the previous step.

  1. Upload the DKU and the Flame family product tar files that you plan on installing to an S3 bucket.

    Note:

    You can also upload the tar files to the instance directly with scp later, but using an S3 bucket means you won't have to re-upload the files again should you need them later.

  2. Launch an instance from the base AMI created in the previous section. You need an instance with a GPU to be able to install the NVIDIA GRID drivers; for this purpose, any G4dn or G5 instance will do.

  3. From your local machine, connect to the instance through ssh. Use the flameadmin account and the keypair that you selected when you finalized the VM settings.

    ssh -i <keypair> flameadmin@<INSTANCE PUBLIC IP>

  4. Create a temporary work folder for later use:

    mkdir -p /tmp/provisioning

  5. Apply some basic configuration settings:

    sudo sed -ir 's/SELINUX=\(disabled\|enforcing\|permissive\)/SELINUX=disabled/' /etc/selinux/config
sudo sed -i 's/GRUB_CMDLINE_LINUX="[^"]*/& net.ifnames=0/' /etc/default/grub
sudo sed -i '/^GRUB_TERMINAL_OUTPUT=/d' /etc/default/grub

sudo tee --append /etc/default/grub << EOF
GRUB_TERMINAL="console serial"
GRUB_SERIAL_COMMAND="serial --speed=115200"
EOF

  6. Set time-related features.

    sudo timedatectl set-local-rtc 0

echo "server 169.254.169.123 prefer iburst minpoll 4 maxpoll 4" | sudo tee --append /etc/chrony.conf

sudo systemctl start chronyd
sudo systemctl enable chronyd

  7. Get the list of time zones with the following command.

    timedatectl list-timezones

  8. From the list of time zones, locate the one you need such as America/New_York and use it to set the time zone on the instance.

    sudo timedatectl set-timezone <Your-Timezone>

  9. Install required tools.

    cd /tmp/provisioning

sudo yum install wget gcc kernel-devel-$(uname -r) -y

wget https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm
sudo rpm -i epel-release-latest-7.noarch.rpm

curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
unzip awscliv2.zip
sudo ./aws/install

  10. Set your credentials for the AWS CLI, based on your secret keypair.

    export AWS_ACCESS_KEY_ID=<YOUR KEY ID>
export AWS_SECRET_ACCESS_KEY=<YOUR SECRET KEY VALUE>
export AWS_DEFAULT_REGION=<YOUR REGION>
    Warning

    Do not use aws configure or store your credentials on the instance, as this would include your credentials in the final AMI!

  11. Install the DKU, replacing <DKU_TAR> with the full path of the tar file uploaded to S3 in step 1.

    cd /tmp/provisioning
aws s3 cp <DKU_TAR> dku.tar
tar xf dku.tar
cd `tar tf dku.tar | head -1 | cut -d "/" -f 1`
sudo ./INSTALL_DKU --silentgeneric

sudo systemctl stop httpd
sudo systemctl disable httpd

  12. Prepare the NVIDIA driver installation.

    sudo tee --append /etc/modprobe.d/blacklist.conf << EOF
blacklist vga16fb
blacklist nouveau
blacklist rivafb
blacklist nvidiafb
blacklist rivatv
EOF

sudo sed -i 's/GRUB_CMDLINE_LINUX="[^"]*/& modprobe.blacklist=nouveau/' /etc/default/grub
sudo grub2-mkconfig -o /boot/grub2/grub.cfg
sudo depmod -a
sudo dracut --force

  13. Reboot the instance:

    sudo reboot

  14. After a few minutes, reconnect to your instance. From your local host:

    ssh -i <keypair> flameadmin@<INSTANCE PUBLIC IP>

  15. Again, set your credentials for the AWS CLI, based on your secret keypair.

    export AWS_ACCESS_KEY_ID=<YOUR KEY ID>
export AWS_SECRET_ACCESS_KEY=<YOUR SECRET KEY VALUE>
export AWS_DEFAULT_REGION=<YOUR REGION>
    Warning

    Do not use aws configure or store your credentials on the instance, as this would include your credentials in the final AMI!

  16. Install the NVIDIA GRID driver.

    cd /tmp/provisioning
aws s3 cp s3://ec2-linux-nvidia-drivers/grid-12.4/NVIDIA-Linux-x86_64-460.106.00-grid-aws.run nvidia-install-script
sudo init 3
sudo /bin/sh nvidia-install-script -s --install-libglvnd
    Note:

    The previous commands download and install NVIDIA GRID drivers version 460.106.00 from AWS S3 storage.

    For more information, see AWS NVIDIA GRID drivers documentation.

  17. Optimize the NVIDIA GPU settings according to your instance type.

    • For a G4dn instance, enter the following settings on the command line.

      sudo nvidia-persistenced
sudo nvidia-smi -ac 5001,1590

    • For a G5 instance, use the following.

      sudo nvidia-persistenced
sudo nvidia-smi -ac 6250,1710
      Note:

      If you launch from this AMI an instance on something other than G4dn, you need to edit these values. To optimize the GPU for instances other than G4dn, see AWS documentation.

  18. Remove the current X config file:

    sudo rm -f /etc/X11/xorg.conf

  19. Create a new /etc/X11/xorg.conf file with the following contents:

    Section "ServerLayout"
    Screen 0    "Screen Default 0" 0 0
#    Screen 1    "Screen Dummy Dual GPU Screen" Relative "Screen Default 0" 0 0

    Identifier "XFree86 Configured"
EndSection

Section "Files"
    FontPath "unix/:7100"
EndSection

Section "Module"
    Load  "glx"    # OpenGL X protocol interface
    Load  "extmod" # Misc. required extensions
EndSection

Section "ServerFlags"
    Option "VTSysReq"       "on"
    Option "DontVTSwitch"   "on"
    Option "DontZoom"       "on"
    Option "AutoAddDevices" "True"
    Option "AutoAddGPU"     "false"
EndSection

Section "InputClass"
    Identifier      "Keyboard0"
    MatchIskeyboard "on"
    Option          "XkbModel"   "pc105"
    Option          "XkbCompat"  "basic+misc"
    Option          "XkbLayout"  "us"
    Option          "XkbOptions" "terminate:ctrl_alt_bksp"
EndSection

Section "InputClass"
    Identifier   "ERASER"
    MatchDriver  "wacom"
    MatchProduct "eraser|Eraser|ERASER"
    Option       "Mode" "Absolute"
EndSection

Section "InputClass"
    Identifier   "cursor"

    MatchDriver  "wacom"
    MatchProduct "cursor|Cursor|CURSOR"
EndSection

Section "InputClass"
    Identifier  "pad"

    MatchDriver  "wacom"
    MatchProduct "pad|Pad|PAD"
    Option       "Mode" "Absolute"
EndSection

Section "InputClass"
    Identifier   "touch"

    MatchDriver  "wacom"
    MatchProduct "touch|Touch|TOUCH|Finger|finger|FINGER"
    Option       "Touch" "off"    
EndSection

Section "Monitor"
    Identifier "Generic EDID"
    VendorName "---"
    ModelName  "NVIDIA Generic EDID"
EndSection

Section "Device"
    Identifier "Dual GPU"
    Driver     "nvidia"
    BoardName  "Unknown"
#Dual GPU    BusId       "PCI:XX:YY:ZZ"
    Option     "UseDisplayDevice" "none"
    Option     "Interactive"      "False"
    Option     "HardDPMS"         "False"
EndSection

Section "Monitor"
    Identifier  "Generic 1920x1200 EDID"
    VendorName  "---"
    ModelName   "NVIDIA Generic 1920x1200 EDID"
    HorizSync   30.0 - 95.0
    VertRefresh 50.0 - 180.0
EndSection

Section "Screen"
    Identifier   "Screen Dummy Dual GPU Screen"
    Device       "Dual GPU"
    Monitor      "Generic 1920x1200 EDID"
    DefaultDepth 24
    SubSection   "Display"
      Virtual 1920 1200
      Depth   24
    EndSubSection
EndSection

Section "Monitor"
    Identifier "Generic Monitor"
EndSection

Section "Device"
    Identifier "NVIDIA Generic"
    Driver     "nvidia"
    Option     "Overlay"        "on"
    Option     "HardDPMS"       "False"
    Option     "Interactive"    "False"
    Option     "RegistryDwords" "PowerMizerEnable=0x1; PerfLevelSrc=0x2222"
EndSection

Section "Screen"
    Identifier   "Screen Default 0"
    Device       "NVIDIA Generic"
    Monitor      "Generic Monitor"
    DefaultDepth 24
    SubSection   "Display"
    Depth        24
    Option       "FlatPanelProperties" "Dithering=Disabled"
    EndSubSection
EndSection

  20. Install the Wacom driver:

    cd /tmp/provisioning

wget https://github.com/linuxwacom/input-wacom/releases/download/input-wacom-0.45.0/input-wacom-0.45.0.tar.bz2
tar xf input-wacom-0.45.0.tar.bz2
cd input-wacom-0.45.0
./configure
make && sudo make install

sudo tee --append /etc/X11/xorg.conf.d/99-wacom-pressure2k.conf << EOF
Section "InputClass"
    Identifier "Wacom pressure compatibility"
    MatchDriver "wacom"
    Option "Pressure2K" "true"
EndSection
EOF

  21. Install Flame family applications, replacing the <... TAR FILE> tags with the actual paths of the files uploaded in step 1.

    • If installing Flame.

      cd /tmp/provisioning

aws s3 cp <FLAME TAR FILE> flame.tar
tar xf flame.tar
cd `tar tf flame.tar | head -1 | cut -d "/" -f 1`
sudo ./INSTALL_FLAME --keepxorg --noui --noagreement

sudo sed -ir 's/^Audiodevice ALSA/#Audiodevice ALSA/' /opt/Autodesk/flame*/cfg/init.cfg
sudo sed -ir 's/^#Audiodevice PulseAudio/Audiodevice PulseAudio/' /opt/Autodesk/flame*/cfg/init.cfg

echo "export DL_VM_ENVIRONMENT=true" | sudo tee --append /etc/environment

    • If installing Flare.

      cd /tmp/provisioning

aws s3 cp <FLARE TAR FILE> flare.tar
tar xf flare.tar
cd `tar tf flare.tar | head -1 | cut -d "/" -f 1`
sudo ./INSTALL_FLARE --keepxorg --noui --noagreement

sudo sed -ir 's/^Audiodevice ALSA/#Audiodevice ALSA/' /opt/Autodesk/flare*/cfg/init.cfg
sudo sed -ir 's/^#Audiodevice PulseAudio/Audiodevice PulseAudio/' /opt/Autodesk/flare*/cfg/init.cfg

echo "export DL_VM_ENVIRONMENT=true" | sudo tee --append /etc/environment

    • If installing Flame Assist.

      cd /tmp/provisioning

aws s3 cp <FLAME_ASSIST TAR FILE> flameassist.tar
tar xf flameassist.tar
cd `tar tf flameassist.tar | head -1 | cut -d "/" -f 1`
sudo ./INSTALL_FLAMEASSIST --keepxorg --noui --noagreement

sudo sed -ir 's/^Audiodevice ALSA/#Audiodevice ALSA/' /opt/Autodesk/flameassist*/cfg/init.cfg
sudo sed -ir 's/^#Audiodevice PulseAudio/Audiodevice PulseAudio/' /opt/Autodesk/flameassist*/cfg/init.cfg

echo "export DL_VM_ENVIRONMENT=true" | sudo tee --append /etc/environment

  22. Clean up:

    sudo sed -i 's/UUID=.*$/UUID=/' /opt/Autodesk/cfg/network.cfg
sudo sed -i 's/ID=.*$/ID=/' /opt/Autodesk/sw/cfg/sw_storage.cfg
dbus-uuidgen | sudo tee /etc/machine-id
sudo find /opt/Autodesk/ -type f \( -name '*.log' -o -name '*.log.[1-9]*' \) -exec rm -f {} \;
sudo sed -i 's|^\([# \\t]*\)Scope=224.0.0.1|Scope=239.0.0.1|' /opt/Autodesk/cfg/network.cfg
sudo sed -i 's/224.0.0.1/239.0.0.1/g' /opt/Autodesk/*/bin/verifyBurnConn

echo "net.ipv4.conf.all.force_igmp_version=2" | sudo tee --append /etc/sysctl.conf

sudo shred -u ~/.aws* /etc/ssh/*_key /etc/ssh/*_key.pub ~/.*history
cd && sudo rm -fr /tmp/provisioning

  23. Clear the command line history and shut down the instance.

    history -c
sudo shutdown now

You can now convert this instance image to a Flame AMI.

Creating the final AMI

  1. From the AWS Console, go the EC2 Instances page.
  2. Right-click on the stopped instance that was created in the previous section.
  3. Select Image and Templates > Create Image.
  4. Give the new image a name and a description.
  5. Select No reboot.
  6. Decide if you want the instance volume to be deleted on termination (Enabled by default).
  7. You can keep all other options as default, or adjust according to your needs.
  8. Click Create Image.
  9. Your new AMI will appear in the AMI section of the EC2 Management Console.

You can now use this AMI to launch Flame-ready instances.