Remove SGs from non-existent nodes ports on Update

[EmilienM]

Current code assumes that if a Node has been removed from the list on UpdateLoadBalancer(), we don't care about it's security group.
We could update applyNodeSecurityGroupIDForLB() to also lookup ports that have the SG tag and detach the SG from the ports that aren't belonging to the ports of the current list of Nodes.

cloud-provider-openstack/pkg/openstack/loadbalancer_sg.go

Lines 72 to 99 in 04bd8e1

	for _, port := range allPorts {
	// You can't assign an SG to a port with port_security_enabled=false, skip them.
	if !port.PortSecurityEnabled {
	continue
	}
	// If the Security Group is already present on the port, skip it.
	if slices.Contains(port.SecurityGroups, sg) {
	continue
	}
	// Only add SGs to the port actually attached to the LB
	if !isPortMember(port, addr, svcConf.lbMemberSubnetID) {
	continue
	}
	// Add the SG to the port
	// TODO(dulek): This isn't an atomic operation. In order to protect from lost update issues we should use
	// `revision_number` handling to make sure our update to `security_groups` field wasn't preceded
	// by a different one. Same applies to a removal of the SG.
	newSGs := append(port.SecurityGroups, sg)
	updateOpts := neutronports.UpdateOpts{SecurityGroups: &newSGs}
	mc := metrics.NewMetricContext("port", "update")
	res := neutronports.Update(context.TODO(), network, port.ID, updateOpts)
	if mc.ObserveRequest(res.Err) != nil {
	return fmt.Errorf("failed to update security group for port %s: %v", port.ID, res.Err)
	}
	}

[EmilienM]

/cc dulek mdbooth

@dulek hey there, you filed a bug some time ago about this issue and I'm taking over it. If time permits, let me know what you think about the proposed fix.

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale