Move files out of /tmp

Signed-off-by: Hanwen <[email protected]>

Author: hanwen-cluster

cookbooks/aws-parallelcluster-environment/recipes/config/network_interfaces.rb

@@ -66,7 +66,7 @@ def cidr_to_netmask(cidr)
 
   cookbook_file 'configure_nw_interface.sh' do
     source 'network_interfaces/configure_nw_interface.sh'
-    path '/tmp/configure_nw_interface.sh'
+    path "#{node['cluster']['tmp_dir']}/configure_nw_interface.sh"
     user 'root'
     group 'root'
     mode '0644'
@@ -87,7 +87,7 @@ def cidr_to_netmask(cidr)
     execute 'configure_nw_interface' do
       user 'root'
       group 'root'
-      cwd "/tmp"
+      cwd node['cluster']['tmp_dir']
       environment(
         # TODO: The variables are a superset of what's required by individual scripts. Consider simplification.
         'DEVICE_NAME' => device_name,
@@ -101,7 +101,7 @@ def cidr_to_netmask(cidr)
         'MAC' => mac
       )
 
-      command 'sh /tmp/configure_nw_interface.sh'
+      command "sh #{node['cluster']['tmp_dir']}/configure_nw_interface.sh"
     end
   end
 

cookbooks/aws-parallelcluster-environment/recipes/install/cfn_bootstrap.rb

@@ -65,7 +65,7 @@
   bucket = "s3.#{aws_region}.#{aws_domain}"
 end
 
-remote_file "/tmp/#{cfnbootstrap_package}" do
+remote_file "#{node['cluster']['tmp_dir']}/#{cfnbootstrap_package}" do
   source "https://#{bucket}/cloudformation-examples/#{cfnbootstrap_package}"
   retries 3
   retry_delay 5
@@ -74,7 +74,7 @@
 bash "Install CloudFormation helpers from #{cfnbootstrap_package}" do
   user 'root'
   group 'root'
-  cwd '/tmp'
+  cwd node['cluster']['tmp_dir']
   code "#{virtualenv_path}/bin/pip install #{cfnbootstrap_package}"
   creates "#{virtualenv_path}/bin/cfn-hup"
 end

cookbooks/aws-parallelcluster-environment/spec/unit/recipes/cfn_bootstrap_spec.rb

@@ -46,7 +46,7 @@
         end
 
         it 'downloads cfn_bootstrap package from s3' do
-          is_expected.to create_remote_file("/tmp/#{cfnbootstrap_package}").with(
+          is_expected.to create_remote_file("#{node['cluster']['tmp_dir']}/#{cfnbootstrap_package}").with(
             source: "https://s3.amazonaws.com/cloudformation-examples/#{cfnbootstrap_package}"
           )
         end
@@ -55,7 +55,7 @@
           is_expected.to run_bash("Install CloudFormation helpers from #{cfnbootstrap_package}").with(
             user: 'root',
             group: 'root',
-            cwd: '/tmp',
+            cwd: node['cluster']['tmp_dir'],
             code: "#{virtualenv_path}/bin/pip install #{cfnbootstrap_package}",
             creates: "#{virtualenv_path}/bin/cfn-hup"
           )
@@ -111,7 +111,7 @@
           runner.converge(described_recipe)
         end
         it 'downloads cfn_bootstrap package from a different s3 bucket' do
-          is_expected.to create_remote_file("/tmp/#{cfnbootstrap_package}").with(
+          is_expected.to create_remote_file("#{node['cluster']['tmp_dir']}/#{cfnbootstrap_package}").with(
             source: "https://s3.cn-north-1.amazonaws.com.cn/cn-north-1-aws-parallelcluster/cloudformation-examples/#{cfnbootstrap_package}"
           )
         end

cookbooks/aws-parallelcluster-environment/test/controls/network_interfaces_spec.rb

@@ -12,7 +12,7 @@
 control 'network_interfaces_configuration_script_created' do
   title 'Check script to configure network interface is created'
 
-  describe file('/tmp/configure_nw_interface.sh') do
+  describe file("#{node['cluster']['tmp_dir']}/configure_nw_interface.sh") do
     it { should exist }
     its('mode') { should cmp '0644' }
     its('owner') { should eq 'root' }

cookbooks/aws-parallelcluster-platform/files/ami_cleanup.sh

@@ -4,6 +4,7 @@
 cloud-init clean -s
 
 rm -rf /var/tmp/* /tmp/*
+rm -rf node['cluster']['tmp_dir']
 rm -rf /etc/ssh/ssh_host_*
 rm -f /etc/udev/rules.d/70-persistent-net.rules
 grep -l "Created by cloud-init on instance boot automatically" /etc/sysconfig/network-scripts/ifcfg-* | xargs rm -f

cookbooks/aws-parallelcluster-platform/recipes/install/cuda.rb

@@ -27,8 +27,8 @@
 cuda_url = "#{node['cluster']['artifacts_s3_url']}/dependencies/cuda/cuda_#{cuda_complete_version}_#{cuda_version_suffix}_#{cuda_arch}.run"
 cuda_samples_version = '12.4'
 cuda_samples_url = "#{node['cluster']['artifacts_s3_url']}/dependencies/cuda/samples/v#{cuda_samples_version}.tar.gz"
-tmp_cuda_run = '/tmp/cuda.run'
-tmp_cuda_sample_archive = '/tmp/cuda-sample.tar.gz'
+tmp_cuda_run = "#{node['cluster']['tmp_dir']}/cuda.run"
+tmp_cuda_sample_archive = "#{node['cluster']['tmp_dir']}/cuda-sample.tar.gz"
 
 node.default['cluster']['nvidia']['cuda']['version'] = cuda_version
 node.default['cluster']['nvidia']['cuda_samples_version'] = cuda_samples_version
@@ -47,13 +47,13 @@
 bash 'cuda.run advanced' do
   user 'root'
   group 'root'
-  cwd '/tmp'
+  cwd node['cluster']['tmp_dir']
   code <<-CUDA
     set -e
     mkdir /cuda-install
     ./cuda.run --silent --toolkit --samples --tmpdir=/cuda-install
     rm -rf /cuda-install
-    rm -f /tmp/cuda.run
+    rm -f #{node['cluster']['tmp_dir']}/cuda.run
   CUDA
   creates "/usr/local/cuda-#{cuda_version}"
 end
@@ -71,11 +71,11 @@
 bash 'cuda.sample install' do
   user 'root'
   group 'root'
-  cwd '/tmp'
+  cwd node['cluster']['tmp_dir']
   code <<-CUDA
     set -e
-    tar xf "/tmp/cuda-sample.tar.gz" --directory "/usr/local/"
-    rm -f "/tmp/cuda-sample.tar.gz"
+    tar xf "#{node['cluster']['tmp_dir']}/cuda-sample.tar.gz" --directory "/usr/local/"
+    rm -f "#{node['cluster']['tmp_dir']}/cuda-sample.tar.gz"
   CUDA
   creates "/usr/local/cuda-#{cuda_version}/samples"
 end

cookbooks/aws-parallelcluster-platform/resources/nvidia_driver/partial/_nvidia_driver_common.rb

@@ -17,7 +17,7 @@
 
 property :nvidia_driver_version, String
 
-tmp_nvidia_run = '/tmp/nvidia.run'
+tmp_nvidia_run = "#{node['cluster']['tmp_dir']}/nvidia.run"
 
 action :setup do
   return unless nvidia_driver_enabled?
@@ -76,11 +76,11 @@
   bash 'nvidia.run advanced' do
     user 'root'
     group 'root'
-    cwd '/tmp'
+    cwd node['cluster']['tmp_dir']
     code <<-NVIDIA
       set -e
       #{compiler_path} ./nvidia.run --silent --dkms --disable-nouveau -m=#{nvidia_kernel_module}
-      rm -f /tmp/nvidia.run
+      rm -f #{node['cluster']['tmp_dir']}/nvidia.run
     NVIDIA
     creates '/usr/bin/nvidia-smi'
   end

cookbooks/aws-parallelcluster-platform/spec/unit/recipes/cuda_spec.rb

@@ -39,7 +39,7 @@
     end
 
     it 'downloads CUDA run file' do
-      is_expected.to create_remote_file('/tmp/cuda.run').with(
+      is_expected.to create_remote_file("#{node['cluster']['tmp_dir']}/cuda.run").with(
         source: cuda_url,
         mode: '0755',
         retries: 3,
@@ -53,16 +53,16 @@
         .with(
           user: 'root',
           group: 'root',
-          cwd: '/tmp',
+          cwd: node['cluster']['tmp_dir'],
           creates: "/usr/local/cuda-#{cuda_version}")
         .with_code(%r{mkdir /cuda-install})
         .with_code(%r{./cuda.run --silent --toolkit --samples --tmpdir=/cuda-install})
         .with_code(%r{rm -rf /cuda-install})
-        .with_code(%r{rm -f /tmp/cuda.run})
+        .with_code(%r{rm -f #{node['cluster']['tmp_dir']}/cuda.run})
     end
 
     it 'downloads CUDA sample files' do
-      is_expected.to create_remote_file('/tmp/cuda-sample.tar.gz').with(
+      is_expected.to create_remote_file("#{node['cluster']['tmp_dir']}/cuda-sample.tar.gz").with(
         source: cuda_samples_url,
         mode: '0644',
         retries: 3,
@@ -75,9 +75,9 @@
         .with(
           user: 'root',
           group: 'root',
-          cwd: '/tmp')
-        .with_code(%r{tar xf "/tmp/cuda-sample.tar.gz" --directory "/usr/local/"})
-        .with_code(%r{rm -f "/tmp/cuda-sample.tar.gz"})
+          cwd: node['cluster']['tmp_dir'])
+        .with_code(%r{tar xf "#{node['cluster']['tmp_dir']}/cuda-sample.tar.gz" --directory "/usr/local/"})
+        .with_code(%r{rm -f "#{node['cluster']['tmp_dir']}/cuda-sample.tar.gz"})
     end
   end
 
@@ -95,7 +95,7 @@
     cached(:node) { chef_run.node }
 
     it 'downloads CUDA run file' do
-      is_expected.to create_remote_file('/tmp/cuda.run').with_source(cuda_url)
+      is_expected.to create_remote_file("#{node['cluster']['tmp_dir']}/cuda.run").with_source(cuda_url)
     end
   end
 end

cookbooks/aws-parallelcluster-platform/spec/unit/resources/nvidia_driver_spec.rb

@@ -198,7 +198,7 @@ def self.setup(chef_run, nvidia_driver_version: nil)
         end
 
         it 'downloads nvidia driver' do
-          is_expected.to create_remote_file('/tmp/nvidia.run').with(
+          is_expected.to create_remote_file("#{node['cluster']['tmp_dir']}/nvidia.run").with(
             source: nvidia_driver_url,
             mode: '0755',
             retries: 3,
@@ -241,11 +241,11 @@ def self.setup(chef_run, nvidia_driver_version: nil)
               .with(
                 user: 'root',
                 group: 'root',
-                cwd: '/tmp',
+                cwd: node['cluster']['tmp_dir'],
                 creates: '/usr/bin/nvidia-smi'
               )
               .with_code(%r{CC=/usr/bin/gcc10-gcc ./nvidia.run --silent --dkms --disable-nouveau -m=#{kernel_module}})
-              .with_code(%r{rm -f /tmp/nvidia.run})
+              .with_code(%r{rm -f #{node['cluster']['tmp_dir']}/nvidia.run})
           end
         elsif platform == 'ubuntu' && version == '22.04'
           it 'installs gcc' do
@@ -269,7 +269,7 @@ def self.setup(chef_run, nvidia_driver_version: nil)
               .with(
                                user: 'root',
                                group: 'root',
-                               cwd: '/tmp',
+                               cwd: node['cluster']['tmp_dir'],
                                creates: '/usr/bin/nvidia-smi'
                              )
               .with_code(%r{#{compiler_path} ./nvidia.run --silent --dkms --disable-nouveau -m=#{kernel_module}})
@@ -283,11 +283,11 @@ def self.setup(chef_run, nvidia_driver_version: nil)
               .with(
                 user: 'root',
                 group: 'root',
-                cwd: '/tmp',
+                cwd: node['cluster']['tmp_dir'],
                 creates: '/usr/bin/nvidia-smi'
               )
               .with_code(%r{./nvidia.run --silent --dkms --disable-nouveau -m=#{kernel_module}})
-              .with_code(%r{rm -f /tmp/nvidia.run})
+              .with_code(%r{rm -f #{node['cluster']['tmp_dir']}/nvidia.run})
           end
         end
 

cookbooks/aws-parallelcluster-shared/attributes/cluster.rb

@@ -1,6 +1,7 @@
 default['cluster']['base_dir'] = '/opt/parallelcluster'
 default['cluster']['sources_dir'] = "#{node['cluster']['base_dir']}/sources"
 default['cluster']['scripts_dir'] = "#{node['cluster']['base_dir']}/scripts"
+default['cluster']['tmp_dir'] = "#{node['cluster']['base_dir']}/tmp"
 default['cluster']['license_dir'] = "#{node['cluster']['base_dir']}/licenses"
 default['cluster']['configs_dir'] = "#{node['cluster']['base_dir']}/configs"
 default['cluster']['shared_dir'] = "#{node['cluster']['base_dir']}/shared"