From fe4e7a866cf40938e6c06201072a83e0ad565405 Mon Sep 17 00:00:00 2001
From: Patrick Dillon <padillon@redhat.com>
Date: Mon, 1 Jun 2026 19:30:12 -0400
Subject: [PATCH] GCP Destroy: Group Load Balancer Resources & Retry on Errors

Groups GCP Load Balancer resources together so they are destroyed
before attempting to destroy subsequent, dependent resources.

Retry delete operations when an internal error is encountered
instead of continually checking the operation with an internal error.
---
 pkg/destroy/gcp/gcp.go | 35 +++++++++++++++++++++++++++++------
 1 file changed, 29 insertions(+), 6 deletions(-)

diff --git a/pkg/destroy/gcp/gcp.go b/pkg/destroy/gcp/gcp.go
index 6a60bc46233..79083173c13 100644
--- a/pkg/destroy/gcp/gcp.go
+++ b/pkg/destroy/gcp/gcp.go
@@ -202,6 +202,17 @@ func (o *ClusterUninstaller) destroyCluster() (bool, error) {
 		{name: "Cloud controller resources", execute: o.discoverCloudControllerResources},
 	}, {
 		{name: "Instances", execute: o.destroyInstances},
+	}, {
+		// LB frontend: forwarding rules reference target TCP proxies and target pools.
+		{name: "Forwarding rules", execute: o.destroyForwardingRules},
+		{name: "Target Pools", execute: o.destroyTargetPools},
+		{name: "Target TCP Proxies", execute: o.destroyTargetTCPProxies},
+	}, {
+		// LB backend: backend services reference instance groups and health checks.
+		{name: "Backend services", execute: o.destroyBackendServices},
+		{name: "Health checks", execute: o.destroyHealthChecks},
+		{name: "HTTP Health checks", execute: o.destroyHTTPHealthChecks},
+	}, {
 		{name: "Disks", execute: o.destroyDisks},
 		{name: "Service accounts", execute: o.destroyServiceAccounts},
 		{name: "Images", execute: o.destroyImages},
@@ -210,13 +221,7 @@ func (o *ClusterUninstaller) destroyCluster() (bool, error) {
 		{name: "Routes", execute: o.destroyRoutes},
 		{name: "Firewalls", execute: o.destroyFirewalls},
 		{name: "Addresses", execute: o.destroyAddresses},
-		{name: "Forwarding rules", execute: o.destroyForwardingRules},
-		{name: "Target Pools", execute: o.destroyTargetPools},
 		{name: "Instance groups", execute: o.destroyInstanceGroups},
-		{name: "Target TCP Proxies", execute: o.destroyTargetTCPProxies},
-		{name: "Backend services", execute: o.destroyBackendServices},
-		{name: "Health checks", execute: o.destroyHealthChecks},
-		{name: "HTTP Health checks", execute: o.destroyHTTPHealthChecks},
 		{name: "Routers", execute: o.destroyRouters},
 		{name: "Subnetworks", execute: o.destroySubnetworks},
 		{name: "Networks", execute: o.destroyNetworks},
@@ -443,6 +448,18 @@ func isErrorStatus(code int64) bool {
 	return code != 0 && (code < 200 || code >= 300)
 }
 
+func hasInternalError(op *compute.Operation) bool {
+	if op.Error == nil {
+		return false
+	}
+	for _, e := range op.Error.Errors {
+		if e.Code == "INTERNAL_ERROR" {
+			return true
+		}
+	}
+	return false
+}
+
 func operationErrorMessage(op *compute.Operation) string {
 	errs := []string{}
 	if op.Error != nil {
@@ -479,6 +496,12 @@ func (o *ClusterUninstaller) handleOperation(ctx context.Context, op *compute.Op
 		o.resetRequestID(identifier...)
 		return fmt.Errorf("failed to delete %s %s with error: %s", resourceType, item.name, operationErrorMessage(op))
 	}
+	if op != nil && op.Status != DONE && hasInternalError(op) {
+		o.resetRequestID(identifier...)
+		o.errorTracker.suppressWarning(item.key, fmt.Errorf("GCP operation %s for %s %s hit an internal error and will be retried: %s",
+			op.Name, resourceType, item.name, operationErrorMessage(op)), o.Logger)
+		return fmt.Errorf("delete %s %s: GCP internal error, retrying", resourceType, item.name)
+	}
 	if (err != nil && isNoOp(err)) || (op != nil && op.Status == DONE) {
 		o.resetRequestID(identifier...)
 		o.deletePendingItems(item.typeName, []cloudResource{item})