What does this PR do?

Fixes ryuk recreation such that a new container is spun up if the previous container has terminated and been removed. This involves checking for ErrNotFound during the Ryuk state check, triggering the creation of a new ryuk container if this error is returned.

It is essential to reset reaperOnce. Without resetting, no attempt to create a new reaper will be made.

Why is it important?

• Test suites that exceed the ryuk timeout fail because a new container is not created.
• Relying on clients to extend timeout as a workaround adds unnecessary work for end users.

Related issues

• Closes [Bug]: ryuk restart/create after first instance has timed out fails #1671

How to test this PR

Run test Test_RecreateReaperIfTerminated to validate changes.

Or run the following test suite with env variable SLEEP=11s. Without changes made in this PR TestC would fail, but with them all tests pass.

package foo

import (
	"context"
	"os"
	"testing"
	"time"

	"github.com/testcontainers/testcontainers-go/modules/redpanda"
)

func TestA(t *testing.T) {
	ctx := context.Background()
	tc, err := redpanda.RunContainer(ctx)
	if err != nil {
		t.Fatal(err)
	}
	t.Cleanup(func() { tc.Terminate(ctx) })
}

func TestB(t *testing.T) {
	d := os.Getenv("SLEEP")
	if d == "" {
		return
	}
	dur, err := time.ParseDuration(d)
	if err != nil {
		t.Fatal(err)
	}
	time.Sleep(dur)
}

func TestC(t *testing.T) {
	ctx := context.Background()
	tc, err := redpanda.RunContainer(ctx)
	if err != nil {
		t.Fatal(err)
	}
	t.Cleanup(func() { tc.Terminate(ctx) })
}