Merge pull request #19555 from lfu/orch_prov_failed_1773595

Improve the error handling of OrchestrationTemplateRunner

(cherry picked from commit 6a79d77)

Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1783389

app/models/manageiq/providers/cloud_manager/orchestration_template_runner.rb

@@ -42,8 +42,8 @@ def deploy_orchestration_stack
     save!
     my_signal(false, :poll_stack_status, 10)
   rescue StandardError => err
-    _log.log_backtrace(err)
-    my_signal(minimize_indirect, :post_stack_run, err.message, 'error')
+    _log.error("Error deploying orchestration stack : #{err.class} - #{err.message}")
+    my_signal(minimize_indirect, :abort_job, err.message, 'error')
   end
 
   def update_orchestration_stack
@@ -55,8 +55,8 @@ def update_orchestration_stack
     save!
     my_signal(false, :poll_stack_status, 10)
   rescue StandardError => err
-    _log.log_backtrace(err)
-    my_signal(minimize_indirect, :post_stack_run, err.message, 'error')
+    _log.error("Error updating orchestration stack : #{err.class} - #{err.message}")
+    my_signal(minimize_indirect, :abort_job, err.message, 'error')
   end
 
   def poll_stack_status(interval)
@@ -72,7 +72,7 @@ def poll_stack_status(interval)
       my_signal(minimize_indirect, :post_stack_run, "Orchestration stack [#{orchestration_stack.name}] #{status}", 'ok')
     when 'rollback_complete', 'delete_complete', /failed$/, /canceled$/
       _log.error("Orchestration stack deployment error: #{message}. Please examine stack resources for more details")
-      my_signal(minimize_indirect, :post_stack_run, "Orchestration stack deployment error: #{message}", 'error')
+      my_signal(minimize_indirect, :abort_job, "Orchestration stack deployment error: #{message}", 'error')
     else
       interval = 60 if interval > 60
       my_signal(false, :poll_stack_status, interval * 2, :deliver_on => Time.now.utc + interval)
@@ -81,8 +81,8 @@ def poll_stack_status(interval)
     # naming convention requires status to end with "failed"
     options.merge!(:orchestration_stack_status => 'check_status_failed', :orchestration_stack_message => err.message)
     save!
-    _log.log_backtrace(err)
-    my_signal(minimize_indirect, :post_stack_run, err.message, 'error')
+    _log.error("Error polling orchestration stack status : #{err.class} - #{err.message}")
+    my_signal(minimize_indirect, :abort_job, err.message, 'error')
   end
 
   def post_stack_run(message, status)

app/models/service_orchestration.rb

@@ -185,6 +185,8 @@ def orchestration_runner_job
 
   def wait_on_orchestration_stack
     while deploy_stack_job.orchestration_stack.blank?
+      raise _("Orchestration template runner finished with error. Check evm.log for details.") if deploy_stack_job.status == 'error'
+
       _log.info("Waiting for the deployment of orchestration stack [#{stack_name}]...")
       sleep 2
       # Code running with Rails QueryCache enabled,

spec/models/service_orchestration_spec.rb

@@ -189,6 +189,13 @@
       end.and_return(job)
       service_with_dialog_options.deploy_orchestration_stack
     end
+
+    it 'raises runtime error when job finishes with pre-existing stack' do
+      job = double(:job, :id => 1, :status => 'error', :orchestration_stack => nil)
+      allow(job).to receive(:signal).with(:start)
+      allow(ManageIQ::Providers::CloudManager::OrchestrationTemplateRunner).to receive(:create_job).and_return(job)
+      expect { service_with_dialog_options.deploy_orchestration_stack }.to raise_error(RuntimeError, /Orchestration template runner finished with error/)
+    end
   end
 
   describe '#update_orchestration_stack' do