fix: Rollback to previous cluster when cluster set fails (#631)

## Description:
<!-- Describe this change, how it works, and the motivation behind it.
-->
When `kurtosis cluster set` fails, we wrongly call `kurtosis cluster
set` on `clusterName` again instead of rolling back to
`clusterPriorToUpdate` as intended.

## Is this change user facing?
YES
<!-- If yes, please add the "user facing" label to the PR -->
<!-- If yes, don't forget to include docs changes where relevant -->

## References (if applicable):
<!-- Add relevant Github Issues, Discord threads, or other helpful
information. -->

cli/cli/commands/cluster/set/set.go

@@ -49,11 +49,6 @@ func run(ctx context.Context, flags *flags.ParsedFlags, args *args.ParsedArgs) e
 		return stacktrace.Propagate(err, "'%s' is not a valid name for Kurtosis cluster", clusterName)
 	}
 
-	engineManager, err := engine_manager.NewEngineManager(ctx)
-	if err != nil {
-		return stacktrace.Propagate(err, "An error occurred creating an engine manager.")
-	}
-
 	clusterUpdateSuccessful := false
 	clusterSettingStore := kurtosis_cluster_setting.GetKurtosisClusterSettingStore()
 	clusterPriorToUpdate, err := clusterSettingStore.GetClusterSetting()
@@ -74,15 +69,19 @@ func run(ctx context.Context, flags *flags.ParsedFlags, args *args.ParsedArgs) e
 		if clusterUpdateSuccessful {
 			return
 		}
-		if err = clusterSettingStore.SetClusterSetting(clusterName); err != nil {
+		if err = clusterSettingStore.SetClusterSetting(clusterPriorToUpdate); err != nil {
 			logrus.Errorf("An error happened updating cluster to '%s'. KUrtosis tried to roll back to the "+
 				"previous value '%s' but the roll back failed. You have to roll back manually running "+
 				"'kurtosis %s %s %s'", clusterName, clusterPriorToUpdate, command_str_consts.ClusterCmdStr,
 				command_str_consts.ClusterSetCmdStr, clusterPriorToUpdate)
 		}
 	}()
-	logrus.Infof("Clustet set to '%s', Kurtosis engine will now be restarted", clusterName)
+	logrus.Infof("Cluster set to '%s', Kurtosis engine will now be restarted", clusterName)
 
+	engineManager, err := engine_manager.NewEngineManager(ctx)
+	if err != nil {
+		return stacktrace.Propagate(err, "An error occurred creating an engine manager.")
+	}
 	// We try to do our best to restart an engine on the same version the current on is on
 	_, engineClientCloseFunc, restartEngineErr := engineManager.RestartEngineIdempotently(ctx, logrus.InfoLevel, noEngineVersion, restartEngineOnSameVersionIfAnyRunning)
 	if restartEngineErr != nil {