diff --git a/installer/cmd/tectonic/main.go b/installer/cmd/tectonic/main.go
index a0f7a0a4d86..85eca882bc3 100644
--- a/installer/cmd/tectonic/main.go
+++ b/installer/cmd/tectonic/main.go
@@ -16,8 +16,9 @@ var (
 	clusterInstallCommand = kingpin.Command("install", "Create a new Tectonic cluster")
 	clusterInstallDirFlag = clusterInstallCommand.Flag("dir", "Cluster directory").Default(".").ExistingDir()
 
-	clusterDestroyCommand = kingpin.Command("destroy", "Destroy an existing Tectonic cluster")
-	clusterDestroyDirFlag = clusterDestroyCommand.Flag("dir", "Cluster directory").Default(".").ExistingDir()
+	clusterDestroyCommand   = kingpin.Command("destroy", "Destroy an existing Tectonic cluster")
+	clusterDestroyDirFlag   = clusterDestroyCommand.Flag("dir", "Cluster directory").Default(".").ExistingDir()
+	clusterDestroyContOnErr = clusterDestroyCommand.Flag("continue-on-error", "Log errors, but attempt to continue cleaning up the cluster.  THIS MAY LEAK RESOURCES, because you may not have enough state left after a partial removal to be able to perform a second destroy.").Default("false").Bool()
 
 	convertCommand    = kingpin.Command("convert", "Convert a tfvars.json to a Tectonic config.yaml")
 	convertConfigFlag = convertCommand.Flag("config", "tfvars.json file").Required().ExistingFile()
@@ -34,7 +35,7 @@ func main() {
 	case clusterInstallCommand.FullCommand():
 		w = workflow.InstallWorkflow(*clusterInstallDirFlag)
 	case clusterDestroyCommand.FullCommand():
-		w = workflow.DestroyWorkflow(*clusterDestroyDirFlag)
+		w = workflow.DestroyWorkflow(*clusterDestroyDirFlag, *clusterDestroyContOnErr)
 	case convertCommand.FullCommand():
 		w = workflow.ConvertWorkflow(*convertConfigFlag)
 	}
diff --git a/installer/pkg/workflow/destroy.go b/installer/pkg/workflow/destroy.go
index c4193e542e6..a77b23653ba 100644
--- a/installer/pkg/workflow/destroy.go
+++ b/installer/pkg/workflow/destroy.go
@@ -22,9 +22,12 @@ const (
 // DestroyWorkflow creates new instances of the 'destroy' workflow,
 // responsible for running the actions required to remove resources
 // of an existing cluster and clean up any remaining artefacts.
-func DestroyWorkflow(clusterDir string) Workflow {
+func DestroyWorkflow(clusterDir string, contOnErr bool) Workflow {
 	return Workflow{
-		metadata: metadata{clusterDir: clusterDir},
+		metadata: metadata{
+			clusterDir: clusterDir,
+			contOnErr:  contOnErr,
+		},
 		steps: []step{
 			readClusterConfigStep,
 			generateTerraformVariablesStep,
diff --git a/installer/pkg/workflow/workflow.go b/installer/pkg/workflow/workflow.go
index 508ebf43d9e..fe0749606b5 100644
--- a/installer/pkg/workflow/workflow.go
+++ b/installer/pkg/workflow/workflow.go
@@ -1,6 +1,9 @@
 package workflow
 
-import "github.com/openshift/installer/installer/pkg/config"
+import (
+	log "github.com/Sirupsen/logrus"
+	"github.com/openshift/installer/installer/pkg/config"
+)
 
 // metadata is the state store of the current workflow execution.
 // It is meant to carry state for one step to another.
@@ -12,6 +15,7 @@ type metadata struct {
 	cluster        config.Cluster
 	configFilePath string
 	clusterDir     string
+	contOnErr      bool
 }
 
 // step is the entrypoint of a workflow step implementation.
@@ -28,11 +32,18 @@ type Workflow struct {
 
 // Execute runs all steps in order.
 func (w Workflow) Execute() error {
+	var firstError error
 	for _, step := range w.steps {
 		if err := step(&w.metadata); err != nil {
-			return err
+			if !w.metadata.contOnErr {
+				return err
+			}
+			if firstError == nil {
+				firstError = err
+			}
+			log.Warn(err)
 		}
 	}
 
-	return nil
+	return firstError
 }