Delete cluster name state file whenever slurm accounting is configured or updated (aws#2994)

* Delete cluster name state file whenever slurm accounting is configured or updated

* Add spec test to cover the deletion of the cluster id state

* Update changelog

Author: hgreebe

CHANGELOG.md

@@ -12,6 +12,7 @@ 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.
+- 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.
 

cookbooks/aws-parallelcluster-slurm/recipes/config/config_slurm_accounting.rb

@@ -76,6 +76,10 @@
   action action
 end unless on_docker?
 
+file "/var/spool/slurm.state/clustername" do
+  action "delete"
+end
+
 if node['cluster']['slurmdbd_service_enabled'] == "true"
   # After starting slurmdbd the database may not be fully responsive yet and
   # its bootstrapping may fail. We need to wait for sacctmgr to successfully

cookbooks/aws-parallelcluster-slurm/recipes/update/clear_slurm_accounting.rb

@@ -23,3 +23,7 @@
   supports restart: false
   action %i(disable stop)
 end
+
+file '/var/spool/slurm.state/clustername' do
+  action :delete
+end

cookbooks/aws-parallelcluster-slurm/spec/unit/recipes/clear_slurm_accounting_spec.rb

@@ -0,0 +1,28 @@
+require 'spec_helper'
+
+describe 'aws-parallelcluster-slurm::clear_slurm_accounting' do
+  for_all_oses do |platform, version|
+    context "on #{platform}#{version}" do
+      cached(:chef_run) do
+        runner = runner(platform: platform, version: version) do |node|
+          mock_file_exists("/var/spool/slurm.state/clustername", true)
+          node.override['cluster']['slurmdbd_service_enabled'] = true
+        end
+        runner.converge(described_recipe)
+      end
+      cached(:node) { chef_run.node }
+
+      it 'stops the slurm database daemon' do
+        is_expected.to disable_service("slurmdbd")
+      end
+
+      it 'deletes the Slurm database password update script' do
+        is_expected.to delete_file("#{node['cluster']['scripts_dir']}/slurm/update_slurm_database_password.sh")
+      end
+
+      it 'Removes existing cluster name state file' do
+        is_expected.to delete_file('/var/spool/slurm.state/clustername')
+      end
+    end
+  end
+end

cookbooks/aws-parallelcluster-slurm/spec/unit/recipes/config_slurm_accounting_spec.rb

@@ -10,6 +10,7 @@
               allow_any_instance_of(Object).to receive(:are_mount_or_unmount_required?).and_return(false)
               allow_any_instance_of(Object).to receive(:dig).and_return(true)
               RSpec::Mocks.configuration.allow_message_expectations_on_nil = true
+              mock_file_exists("/var/spool/slurm.state/clustername", true)
               node.override['cluster']['slurmdbd_service_enabled'] = enable_service
             end
             runner.converge(described_recipe)
@@ -70,6 +71,9 @@
             )
           end
           if enable_service == "true"
+            it 'Removes existing cluster name state file' do
+              is_expected.to delete_file('/var/spool/slurm.state/clustername')
+            end
             it 'starts the slurm database daemon' do
               is_expected.to enable_service("slurmdbd")
               is_expected.to start_service("slurmdbd")