Do not block shard ownership assertion if `acquireShards` is blocked

Do not block shard ownership assertion if acquireShards is blocked
v1.17.x logic:

shard controller will start a background thread periodically asserting shard ownership

for {
	select {
	...
	case <-acquireTicker.C:
		c.acquireShards()
	...
	}
}

if acquireShards is blocked due to whatever reason, then no further acquireShards will be executed.

OSS team should consider the following logic

var (
	ShardController struct {
		...

		shardAssertionInProgressMutex sync.Mutex
		shardAssertionInProgress map[int32]struct{}
		shardIDChan chan int32
		...
	}
)

func (sc *ShardController) eventLoop() {
	acquireTicker := time.NewTicker(c.config.AcquireShardInterval())
	defer acquireTicker.Stop()

	for {
		select {
		case <-c.shutdownCh:			
			return
		case <-acquireTicker.C:
			for i = 0; i < totalShardIDs; i++ {
				shardAssertionInProgressMutex.Lock()
				_, exist = sc.shardAssertionInProgress[i]
				if !exist {
					sc.shardAssertionInProgress[i] = struct{}{}
				}
				shardAssertionInProgressMutex.Unlock()
				if !exist {
					sc.shardIDChan <- i
				}
			}
		}
	}
}

func (sc *ShardController) acquireShard() {
	for shardID := range shardIDChan {
		...
		// shard ownership assertion logic here
		...
		
		shardAssertionInProgressMutex.Lock()
		delete(sc.shardAssertionInProgress, shardID)
		shardAssertionInProgressMutex.Unlock()
	}
}

Locate the shard controller's eventLoop, acquireShards, and acquireShard entry points and read how the periodic ownership assertion is currently dispatched. Confirm the change keeps later shard ownership assertions running when one acquire operation blocks, while preserving shutdown behavior and preventing duplicate assertions for the same shard.