Fix interrupt state to account for failure correctly

If the job hasn't completely failed but some of the pods
have failed and we deleted the job we weren't setting the MOSB
to interrupted as expected. This was because we had a wrong check for
failure state. This patch fixes that.

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

Author: umohnani8

pkg/controller/build/imagebuilder/jobimagebuilder.go

@@ -329,7 +329,7 @@ func (j *jobImageBuilder) validateBuilderType(builder buildrequest.Builder) erro
 func MapJobStatusToBuildStatus(job *batchv1.Job) (mcfgv1.BuildProgress, []metav1.Condition) {
 	// If the job is being deleted and it was not in either a successful or failed state
 	// then the MachineOSBuild should be considered "interrupted"
-	if job.DeletionTimestamp != nil && job.Status.Succeeded == 0 && job.Status.Failed == 0 {
+	if job.DeletionTimestamp != nil && job.Status.Succeeded == 0 && job.Status.Failed < constants.JobMaxRetries+1 {
 		return mcfgv1.MachineOSBuildInterrupted, apihelpers.MachineOSBuildInterruptedConditions()
 	}