[DFSM]Using .login_nodes_keys_sync_file to be used during Init and Update phase of the clusters (#2677)

Co-authored-by: Himani Deshpande <[email protected]>

Author: himani2411

cookbooks/aws-parallelcluster-environment/attributes/environment.rb

@@ -55,6 +55,7 @@
 default['cluster']['shared_dir_compute'] = node['cluster']['shared_dir']
 default['cluster']['shared_dir_head'] = node['cluster']['shared_dir']
 default['cluster']['shared_dir_login'] = node['cluster']['shared_dir_login_nodes']
+default['cluster']['shared_login_nodes_keys_sync_file'] = "#{node['cluster']['shared_dir_login_nodes']}/.login_nodes_keys_sync_file"
 # Since this is a shared directory, it needs to be defined here first instead of in the dependent cookbook for slurm
 default['cluster']['slurm']['install_dir'] = '/opt/slurm'
 

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

@@ -21,7 +21,7 @@
 script_dir = "#{keys_dir}/scripts"
 script_path = "#{script_dir}/keys-manager.sh"
 
-sync_file_path = "#{keys_dir}/.login_nodes_keys_sync_file"
+sync_file_path = "#{node['cluster']['shared_login_nodes_keys_sync_file']}"
 
 case node['cluster']['node_type']
 when 'ComputeFleet'

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

@@ -93,9 +93,9 @@
         end
         cached(:node) { chef_run.node }
 
-        it "waits for cluster config version file" do
-          is_expected.to run_bash("Wait for synchronization file at #{SYNC_FILE} to be written for version #{CLUSTER_CONFIG_VERSION}").with(
-            code: "[[ \"$(cat #{SYNC_FILE})\" == \"#{CLUSTER_CONFIG_VERSION}\" ]] || exit 1",
+        it "Wait for sync file to exist" do
+          is_expected.to run_bash("Wait for synchronization file at #{SYNC_FILE} to exist").with(
+            code: "[[ -e #{SYNC_FILE} ]] || exit 1",
             retries: 30,
             retry_delay: 10,
             timeout: 5

cookbooks/aws-parallelcluster-shared/libraries/helpers.rb

@@ -95,13 +95,12 @@ def write_sync_file(path)
 end
 
 def wait_sync_file(path)
-  # Wait for a synchronization file to be written for the current cluster config version.
+  # Wait for a synchronization file to exist.
   # Synchronization files are used as a synchronization point between cluster nodes
   # to signal that a group of actions have been completed.
-  cluster_config_version = node["cluster"]["cluster_config_version"]
   # Wait for the config version file to contain the current cluster config version.
-  bash "Wait for synchronization file at #{path} to be written for version #{cluster_config_version}" do
-    code "[[ \"$(cat #{path})\" == \"#{cluster_config_version}\" ]] || exit 1"
+  bash "Wait for synchronization file at #{path} to exist" do
+    code "[[ -e #{path} ]] || exit 1"
     retries 30
     retry_delay 10
     timeout 5