Merge branch 'develop' into develop

hgreebe · web-flow · commit aa1a78283c06 · 2025-08-08T11:50:20.000-04:00
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -11,7 +11,20 @@ This file is used to list changes made in each version of the AWS ParallelCluste
 
 **CHANGES**
 - Ubuntu 20.04 is no longer supported.
-- Upgrade Slurm to version 24.11.5.
+- Upgrade Slurm to version 24.11.6 (from 24.05.8).
+- Upgrade EFA installer to 1.42.0 (from 1.41.0).
+  - Efa-driver: efa-2.15.3-1
+  - Efa-config: efa-config-1.18-1
+  - Efa-profile: efa-profile-1.7-1
+  - Libfabric-aws: libfabric-aws-2.1.0-3
+  - Rdma-core: rdma-core-57.0-1
+  - Open MPI: openmpi40-aws-4.1.7-2 and openmpi50-aws-5.0.6-11
+- Upgrade Cinc Client to version to 18.4.12 from 18.2.7.
+- Upgrade NVIDIA driver to version 570.172.08 (from 570.86.15) for all OSs except AL2.
+- Upgrade CUDA Toolkit to version 12.8.1 (from 12.8.0) for all OSs except AL2.
+- Upgrade DCGM to version 4.2.3 (from 3.3.6) for all OSs except AL2.
+- Upgrade Python to 3.12.11 (from 3.12.8) for all OSs except AL2.
+- Upgrade Intel MPI Library to 2021.16.0 (from 2021.13.1).
 - Addressed cluster id mismatch known issue by deleting the file `/var/spool/slurm.state/clustername` before configuring Slurm accounting.
 - Upgrade DCV to version 2024.0-19030.
 - Remove `berkshelf`. All cookbooks are local and do not need `berkshelf` dependency management.
diff --git a/cookbooks/aws-parallelcluster-awsbatch/test/controls/awsbatch_virtualenv_spec.rb b/cookbooks/aws-parallelcluster-awsbatch/test/controls/awsbatch_virtualenv_spec.rb
@@ -13,7 +13,7 @@
 pyenv_dir = "#{base_dir}/pyenv"
 
 control 'tag:install_awsbatch_virtualenv_created' do
-  python_version = os_properties.alinux2? ? '3.9.20' : '3.12.8'
+  python_version = os_properties.alinux2? ? '3.9.20' : '3.12.11'
   title "awsbatch virtualenv should be created on #{python_version}"
   only_if { !os_properties.redhat? }
 
diff --git a/cookbooks/aws-parallelcluster-computefleet/kitchen.computefleet-config.yml b/cookbooks/aws-parallelcluster-computefleet/kitchen.computefleet-config.yml
@@ -31,7 +31,7 @@ suites:
     attributes:
       cluster:
         custom_node_package: https://github.com/aws/aws-parallelcluster-node/archive/develop.tar.gz
-        python-version: 3.12.8
+        python-version: 3.12.11
         node_virtualenv_path: /opt/parallelcluster/pyenv/versions/node_virtualenv
   - name: fleet_status
     run_list:
diff --git a/cookbooks/aws-parallelcluster-environment/attributes/environment.rb b/cookbooks/aws-parallelcluster-environment/attributes/environment.rb
@@ -70,8 +70,8 @@
 
 default['cluster']['head_node_private_ip'] = nil
 
-default['cluster']['efa']['version'] = '1.41.0'
-default['cluster']['efa']['sha256'] = '3506354cdfbe31ff552fe75f5d0d9bb7efd29cf79bd99457347d29c751c38f9f'
+default['cluster']['efa']['version'] = '1.42.0'
+default['cluster']['efa']['sha256'] = '4114fe612905ee05083ae5cb391a00a012510f3abfecc642d86c9a5ae4be9008'
 
 default['cluster']['efs']['version'] = '2.3.1'
 default['cluster']['efs']['sha256'] = 'ced12f82e76f9740476b63f30c49bd76cc00b6375e12a9f5f7ba852635c49e15'
diff --git a/cookbooks/aws-parallelcluster-environment/spec/unit/resources/efa_spec.rb b/cookbooks/aws-parallelcluster-environment/spec/unit/resources/efa_spec.rb
@@ -2,8 +2,8 @@
 
 # parallelcluster default source dir defined in attributes
 source_dir = '/opt/parallelcluster/sources'
-efa_version = '1.41.0'
-efa_checksum = '3506354cdfbe31ff552fe75f5d0d9bb7efd29cf79bd99457347d29c751c38f9f'
+efa_version = '1.42.0'
+efa_checksum = '4114fe612905ee05083ae5cb391a00a012510f3abfecc642d86c9a5ae4be9008'
 
 class ConvergeEfa
   def self.setup(chef_run, efa_version: nil, efa_checksum: nil)
diff --git a/cookbooks/aws-parallelcluster-environment/test/controls/cfn_bootstrap_spec.rb b/cookbooks/aws-parallelcluster-environment/test/controls/cfn_bootstrap_spec.rb
@@ -13,7 +13,7 @@
 pyenv_dir = "#{base_dir}/pyenv"
 
 control 'tag:install_cfnbootstrap_virtualenv_created' do
-  cfn_python_version = os_properties.alinux2? ? '3.9.20' : '3.12.8'
+  cfn_python_version = os_properties.alinux2? ? '3.9.20' : '3.12.11'
   title "cfnbootstrap virtualenv should be created on #{cfn_python_version}"
   only_if { !os_properties.redhat_on_docker? }
 
diff --git a/cookbooks/aws-parallelcluster-platform/attributes/platform.rb b/cookbooks/aws-parallelcluster-platform/attributes/platform.rb
@@ -17,9 +17,10 @@
 # NVidia
 default['cluster']['nvidia']['enabled'] = 'no'
 default['cluster']['nvidia']['driver_version'] = '570.172.08'
-default['cluster']['nvidia']['dcgm_version'] = '3.3.6'
+default['cluster']['nvidia']['dcgm_version'] = '4.2.3-2'
 if platform?('amazon') && node['platform_version'] == "2"
   default['cluster']['nvidia']['driver_version'] = '550.127.08'
+  default['cluster']['nvidia']['dcgm_version'] = '3.3.6-1'
 end
 
 # nvidia-imex
diff --git a/cookbooks/aws-parallelcluster-platform/recipes/install/cuda.rb b/cookbooks/aws-parallelcluster-platform/recipes/install/cuda.rb
@@ -20,9 +20,9 @@
 # Cuda installer from https://developer.nvidia.com/cuda-toolkit-archive
 # Cuda installer naming: cuda_11.8.0_520.61.05_linux
 cuda_version = '12.8'
-cuda_patch = '0'
+cuda_patch = '1'
 cuda_complete_version = "#{cuda_version}.#{cuda_patch}"
-cuda_version_suffix = '570.86.10'
+cuda_version_suffix = '570.124.06'
 cuda_samples_version = '12.8'
 if platform?('amazon') && node['platform_version'] == "2"
   cuda_version = '12.4'
diff --git a/cookbooks/aws-parallelcluster-platform/recipes/install/intel_mpi.rb b/cookbooks/aws-parallelcluster-platform/recipes/install/intel_mpi.rb
@@ -16,7 +16,7 @@
 # limitations under the License.
 
 intelmpi_supported = !arm_instance?
-intelmpi_version = '2021.13'
+intelmpi_version = '2021.16'
 
 node.default['conditions']['intel_mpi_supported'] = intelmpi_supported
 node.default['cluster']['intelmpi']['version'] = intelmpi_version
@@ -25,9 +25,9 @@
 
 return unless intelmpi_supported
 
-intelmpi_full_version = "#{intelmpi_version}.1.769"
+intelmpi_full_version = "#{intelmpi_version}.0.443"
 intelmpi_installation_path = "/opt/intel/mpi/#{intelmpi_version}"
-intelmpi_installer = "l_mpi_oneapi_p_#{intelmpi_full_version}_offline.sh"
+intelmpi_installer = "intel-mpi-#{intelmpi_full_version}_offline.sh"
 intelmpi_installer_path = "#{node['cluster']['sources_dir']}/#{intelmpi_installer}"
 intelmpi_installer_url = "#{node['cluster']['artifacts_s3_url']}/impi/#{intelmpi_installer}"
 intelmpi_qt_version = '6.5.3'
diff --git a/cookbooks/aws-parallelcluster-platform/resources/nvidia_dcgm/partial/_nvidia_dcgm_debian.rb b/cookbooks/aws-parallelcluster-platform/resources/nvidia_dcgm/partial/_nvidia_dcgm_debian.rb
@@ -13,34 +13,45 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 action :install_package do
-  remote_file "#{node['cluster']['sources_dir']}/#{dcgm_package}-#{package_version}.deb" do
-    source "#{dcgm_url}"
-    mode '0644'
-    retries 3
-    retry_delay 5
-    action :create_if_missing
-  end
+  packages_urls_list = if package_version.start_with?("3.")
+                         [dcgm_package]
+                       else
+                         [dcgm4_core_package, dcgm4_package]
+                       end
+  packages_urls_list.each do |package|
+    remote_file "#{node['cluster']['sources_dir']}/#{package}-#{package_version}.deb" do
+      source "#{node['cluster']['artifacts_s3_url']}/dependencies/nvidia_dcgm/#{platform}/#{package}_#{package_version}_#{arch_suffix}.deb"
+      mode '0644'
+      retries 3
+      retry_delay 5
+      action :create_if_missing
+    end
 
-  bash "Install #{dcgm_package}" do
-    user 'root'
-    cwd node['cluster']['sources_dir']
-    code <<-DCGM_INSTALL
-    set -e
-    dpkg -i #{dcgm_package}-#{package_version}.deb
-    DCGM_INSTALL
-    retries 3
-    retry_delay 5
+    bash "Install #{package}" do
+      user 'root'
+      cwd node['cluster']['sources_dir']
+      code <<-DCGM_INSTALL
+      set -e
+      dpkg -i #{package}-#{package_version}.deb
+      DCGM_INSTALL
+      retries 3
+      retry_delay 5
+    end
   end
 end
 
-def dcgm_url
-  "#{node['cluster']['artifacts_s3_url']}/dependencies/nvidia_dcgm/#{platform}/#{dcgm_package}_#{package_version}_#{arch_suffix}.deb"
-end
-
 def dcgm_package
   'datacenter-gpu-manager'
 end
 
+def dcgm4_package
+  "#{dcgm_package}-4-cuda12"
+end
+
+def dcgm4_core_package
+  "#{dcgm_package}-4-core"
+end
+
 def arch_suffix
   arm_instance? ? 'arm64' : 'amd64'
 end
diff --git a/cookbooks/aws-parallelcluster-platform/resources/nvidia_dcgm/partial/_nvidia_dcgm_rhel.rb b/cookbooks/aws-parallelcluster-platform/resources/nvidia_dcgm/partial/_nvidia_dcgm_rhel.rb
@@ -13,34 +13,47 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 action :install_package do
-  remote_file "#{node['cluster']['sources_dir']}/#{dcgm_package}-#{package_version}.rpm" do
-    source "#{dcgm_url}"
-    mode '0644'
-    retries 3
-    retry_delay 5
-    action :create_if_missing
+  if package_version.start_with?("3.")
+    packages_urls_list = [dcgm_package]
+    package_url_separator = "-"
+  else
+    packages_urls_list = [dcgm4_core_package, dcgm4_package]
+    package_url_separator = "."
   end
+  packages_urls_list.each do |package|
+    remote_file "#{node['cluster']['sources_dir']}/#{package}-#{package_version}.rpm" do
+      source "#{node['cluster']['artifacts_s3_url']}/dependencies/nvidia_dcgm/#{platform}/#{package}-#{package_version}#{package_url_separator}#{arch_suffix}.rpm"
+      mode '0644'
+      retries 3
+      retry_delay 5
+      action :create_if_missing
+    end
 
-  bash "Install #{dcgm_package}" do
-    user 'root'
-    cwd node['cluster']['sources_dir']
-    code <<-DCGM_INSTALL
-    set -e
-    yum install -y #{dcgm_package}-#{package_version}.rpm
-    DCGM_INSTALL
-    retries 3
-    retry_delay 5
+    bash "Install #{package}" do
+      user 'root'
+      cwd node['cluster']['sources_dir']
+      code <<-DCGM_INSTALL
+      set -e
+      yum install -y #{package}-#{package_version}.rpm
+      DCGM_INSTALL
+      retries 3
+      retry_delay 5
+    end
   end
 end
 
-def dcgm_url
-  "#{node['cluster']['artifacts_s3_url']}/dependencies/nvidia_dcgm/#{platform}/#{dcgm_package}-#{package_version}-1-#{arch_suffix}.rpm"
-end
-
 def dcgm_package
   'datacenter-gpu-manager'
 end
 
+def dcgm4_package
+  "#{dcgm_package}-4-cuda12"
+end
+
+def dcgm4_core_package
+  "#{dcgm_package}-4-core"
+end
+
 def arch_suffix
   arm_instance? ? 'aarch64' : 'x86_64'
 end
diff --git a/cookbooks/aws-parallelcluster-platform/spec/unit/recipes/cuda_spec.rb b/cookbooks/aws-parallelcluster-platform/spec/unit/recipes/cuda_spec.rb
@@ -2,9 +2,9 @@
 
 describe 'aws-parallelcluster-platform::cuda' do
   cached(:cuda_version) { '12.8' }
-  cached(:cuda_patch) { '0' }
+  cached(:cuda_patch) { '1' }
   cached(:cuda_complete_version) { "#{cuda_version}.#{cuda_patch}" }
-  cached(:cuda_version_suffix) { '570.86.10' }
+  cached(:cuda_version_suffix) { '570.124.06' }
 
   context 'when nvidia not enabled' do
     cached(:chef_run) do
diff --git a/cookbooks/aws-parallelcluster-platform/spec/unit/recipes/intel_mpi_spec.rb b/cookbooks/aws-parallelcluster-platform/spec/unit/recipes/intel_mpi_spec.rb
@@ -20,8 +20,8 @@
   end
 
   it 'fetches intel mpi installer script' do
-    is_expected.to create_remote_file("#{source_dir}/l_mpi_oneapi_p_2021.13.1.769_offline.sh").with(
-      source: "https://#{aws_region}-aws-parallelcluster.s3.#{aws_region}.test_aws_domain/archives/impi/l_mpi_oneapi_p_2021.13.1.769_offline.sh",
+    is_expected.to create_remote_file("#{source_dir}/intel-mpi-2021.16.0.443_offline.sh").with(
+      source: "https://#{aws_region}-aws-parallelcluster.s3.#{aws_region}.test_aws_domain/archives/impi/intel-mpi-2021.16.0.443_offline.sh",
       mode: '0744',
       retries: 3,
       retry_delay: 5
@@ -31,25 +31,25 @@
   it 'installs intel mpi' do
     is_expected.to run_bash('install intel mpi').with(
       cwd: source_dir,
-      creates: '/opt/intel/mpi/2021.13'
-    ).with_code(%r{chmod +x l_mpi_oneapi_p_2021.13.1.769_offline.sh --remove-extracted-files yes -a --silent --eula accept --install-dir /opt/intel})
-                                                .with_code(/rm -f l_mpi_oneapi_p_2021.13.1.769_offline.sh/)
+      creates: '/opt/intel/mpi/2021.16'
+    ).with_code(%r{chmod +x intel-mpi-2021.16.0.443_offline.sh --remove-extracted-files yes -a --silent --eula accept --install-dir /opt/intel})
+                                                .with_code(/rm -f intel-mpi-2021.16.0.443_offline.sh/)
   end
 
   it 'appends intel module file dir to modules config' do
     is_expected.to append_to_config_modules('append intel modules file dir to modules conf')
-      .with_line('/opt/intel/mpi/2021.13/etc/modulefiles/')
+      .with_line('/opt/intel/mpi/2021.16/etc/modulefiles/')
   end
 
   it 'renames intel mpi module' do
     is_expected.to run_execute('rename intel mpi modules file name').with(
-      command: "mv /opt/intel/mpi/2021.13/etc/modulefiles/mpi /opt/intel/mpi/2021.13/etc/modulefiles/intelmpi",
-      creates: '/opt/intel/mpi/2021.13/etc/modulefiles/intelmpi'
+      command: "mv /opt/intel/mpi/2021.16/etc/modulefiles/mpi /opt/intel/mpi/2021.16/etc/modulefiles/intelmpi",
+      creates: '/opt/intel/mpi/2021.16/etc/modulefiles/intelmpi'
     )
   end
 
   it 'adds Qt source file' do
-    is_expected.to create_template("/opt/intel/mpi/2021.13/qt_source_code.txt").with(
+    is_expected.to create_template("/opt/intel/mpi/2021.16/qt_source_code.txt").with(
       source: 'intel_mpi/qt_source_code.erb',
       owner: 'root',
       group: 'root',
diff --git a/cookbooks/aws-parallelcluster-platform/spec/unit/resources/nvidia_dcgm_spec.rb b/cookbooks/aws-parallelcluster-platform/spec/unit/resources/nvidia_dcgm_spec.rb
@@ -170,7 +170,8 @@ def self.setup(chef_run, nvidia_enabled: nil)
             end
           else
             it 'installs datacenter gpu manager' do
-              is_expected.to run_bash('Install datacenter-gpu-manager')
+              is_expected.to run_bash('Install datacenter-gpu-manager-4-core')
+              is_expected.to run_bash('Install datacenter-gpu-manager-4-cuda12')
             end
           end
         end
diff --git a/cookbooks/aws-parallelcluster-platform/test/controls/nvidia_dcgm_spec.rb b/cookbooks/aws-parallelcluster-platform/test/controls/nvidia_dcgm_spec.rb
@@ -14,8 +14,13 @@
     ['yes', true, 'true'].include?(node['cluster']['nvidia']['enabled']) && !instance.custom_ami? &&
       (!os_properties.arm? || !(os_properties.alinux2? || os_properties.centos?))
   end
-
-  describe package('datacenter-gpu-manager') do
-    it { should be_installed }
+  if os_properties.alinux2?
+    describe package('datacenter-gpu-manager') do
+      it { should be_installed }
+    end
+  else
+    describe package('datacenter-gpu-manager-4-cuda12') do
+      it { should be_installed }
+    end
   end
 end
diff --git a/cookbooks/aws-parallelcluster-shared/attributes/versions.rb b/cookbooks/aws-parallelcluster-shared/attributes/versions.rb
@@ -1,5 +1,5 @@
 # Python Version
-default['cluster']['python-version'] = '3.12.8'
+default['cluster']['python-version'] = '3.12.11'
 default['cluster']['python-major-minor-version'] = '3.12'
 if platform?('amazon') && node['platform_version'] == "2"
   default['cluster']['python-version'] = '3.9.20'
diff --git a/cookbooks/aws-parallelcluster-slurm/attributes/versions.rb b/cookbooks/aws-parallelcluster-slurm/attributes/versions.rb
@@ -1,8 +1,8 @@
 # Slurm
-default['cluster']['slurm']['version'] = '24-11-5-1'
+default['cluster']['slurm']['version'] = '24-11-6-1'
 default['cluster']['slurm']['commit'] = ''
 default['cluster']['slurm']['branch'] = ''
-default['cluster']['slurm']['sha256'] = 'e1a5547edd212c38b5e3230a284133f777b32746551f094aaa81cc4af375e332'
+default['cluster']['slurm']['sha256'] = '282708483326f381eb001a14852a1a82e65e18f37b62b7a5f4936c0ed443b600'
 default['cluster']['slurm']['base_url'] = "#{node['cluster']['artifacts_s3_url']}/dependencies/slurm"
 # Munge
 default['cluster']['munge']['munge_version'] = '0.5.16'
