fix based on the flaky finalizers test - finalizers ordering and init

[gkech]

CHANGE DESCRIPTION

Problem:

The following test has been failing randomly for quite some time

Cause:
Short explanation of the root cause of the issue if applicable.

Solution:

Scenario that panics and happens ~50% of the time:

1. stop-watchers iterates first and then the stopExternalWatchers runs (no ToCrunchy) > Patch removes it > cr.PGVector reset to nil
2. delete-ssl iterates next > deleteTLSSecrets > ToCrunchy > *cr.Spec.Extensions.BuiltIn.PGVector with nil pointer > panic

Scenario that doesn't panic:

1. delete-ssl iterates first > ToCrunchy with PGVector=&false (still valid from Default()) > no panic
2. stop-watchers iterates next > no ToCrunchy > no panic

By adding nil guards in ToCrunchy it doesn't panic when the *bool extension fields are nil (which can happen when cr is refreshed from the server after a Patch resets the in-memory defaults). The crunchy cluster's extension fields are plain bool, so unset *bool pointers correctly leave them as false. Also the . runFinalizers iterates over a map — which has random iteration order in Go, so we can make that deterministic.

CHECKLIST

Jira

• Is the Jira ticket created and referenced properly?
• Does the Jira ticket have the proper statuses for documentation (Needs Doc) and QA (Needs QA)?
• Does the Jira ticket link to the proper milestone (Fix Version field)?

Tests

• Is an E2E test/test case added for the new feature/change?
• Are unit tests added where appropriate?

Config/Logging/Testability

• Are all needed new/changed options added to default YAML files?
• Are all needed new/changed options added to the Helm Chart?
• Did we add proper logging messages for operator actions?
• Did we ensure compatibility with the previous version or cluster upgrade process?
• Does the change support oldest and newest supported PG version?
• Does the change support oldest and newest supported Kubernetes version?

[gkech]

small test:

package main

import "fmt"

func main() {
	values := map[string]string{
		"key1": "value1",
		"key2": "value2",
		"key3": "value3",
		"key4": "value4",
		"key5": "value5",
		"key6": "value6",
		"key7": "value7",
		"key8": "value8",
	}

	for key, value := range values {
		fmt.Println(key, value)
	}
}

Prints:

key3 value3
key4 value4
key5 value5
key6 value6
key7 value7
key8 value8
key1 value1
key2 value2

[JNKPercona]

Test Name	Result	Time
backup-enable-disable	passed	00:10:07
builtin-extensions	passed	00:05:36
cert-manager-tls	passed	00:06:04
custom-envs	passed	00:18:24
custom-extensions	passed	00:13:14
custom-tls	passed	00:05:20
database-init-sql	passed	00:02:38
demand-backup	passed	00:24:03
demand-backup-offline-snapshot	passed	00:12:43
dynamic-configuration	passed	00:03:26
finalizers	passed	00:03:54
init-deploy	passed	00:03:05
huge-pages	passed	00:02:45
monitoring	passed	00:07:06
monitoring-pmm3	passed	00:08:05
one-pod	passed	00:05:45
operator-self-healing	passed	00:10:36
pitr	passed	00:14:08
scaling	passed	00:04:48
scheduled-backup	passed	00:24:05
self-healing	passed	00:09:02
sidecars	passed	00:02:42
standby-pgbackrest	passed	00:11:57
standby-streaming	passed	00:09:02
start-from-backup	passed	00:10:35
tablespaces	passed	00:06:41
telemetry-transfer	passed	00:04:33
upgrade-consistency	passed	00:05:35
upgrade-minor	passed	00:06:00
users	passed	00:04:33

Summary	Value
Tests Run	30/30
Job Duration	01:23:38
Total Test Time	04:16:47

commit: 1e36c1b
image: perconalab/percona-postgresql-operator:PR-1460-1e36c1b8e