The broker cannot join the cluster after it is started #2833
Description
我使用开源版 1.5.5 在虚拟机上部署了一个集群，但是节点在初次启动 automq 进程时无法加入集群，终止进程后，再次以同样的命令启动后又正常加入了集群。
如下，我有一个 10.20.0.127 的新节点作为 broker 启动，但是在初次启动时会一直报错, 节点始终无法正常加入集群。当我 kill 进程，再重新启动时，节点就可以正常加入集群了。这两启动之间外部环境和启动命令没有任何变化，这是什么原因呢？broker 节点和 controller节点都存在同样的问题，controller 节点初次启动时也无法组成集群自动退出，再次启动后正常。以下为 broker 节点的日志，附件为 broker 两次启动的完整日志
broker 启动命令：
KAFKA_S3_ACCESS_KEY='XXXXXXXXXXXXXXXXXXXX' KAFKA_S3_SECRET_KEY='XXXXXXXXXXXXXXXXXXXXXXXXXXX' ./bin/kafka-server-start.sh -daemon config/kraft/broker.properties --override cluster.id=sis-automq-test --override node.id=1005 --override controller.quorum.voters=0@10.20.9.210:9093,1@10.20.11.91:9093,2@10.20.1.32:9093 --override controller.quorum.bootstrap.servers=10.20.9.210:9093,10.20.11.91:9093,10.20.1.32:9093 --override advertised.listeners=PLAINTEXT://10.20.0.127:9092 --override s3.data.buckets='0@s3://shareit-tmp-sg?region=ap-southeast-3&endpoint=https://obs.ap-southeast-3.myhuaweicloud.com' --override s3.wal.path='0@s3://shareit-tmp-sg?region=ap-southeast-3&endpoint=https://obs.ap-southeast-3.myhuaweicloud.com' --override s3.ops.buckets='1@s3://shareit-tmp-sg?region=ap-southeast-3&endpoint=https://obs.ap-southeast-3.myhuaweicloud.com' --override log.dirs='/root/kraft-logs' --override s3.telemetry.metrics.exporter.type=prometheus --override s3.metrics.exporter.prom.host=0.0.0.0 --override s3.metrics.exporter.prom.port=10106第一次启动：
[2025-09-12 16:19:17,065] INFO [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Node -1 disconnected. (org.apache.kafka.clients.NetworkClient)
[2025-09-12 16:19:17,066] WARN [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Connection to node -1 (localhost/127.0.0.1:9092) could not be establ
ished. Node may not be available. (org.apache.kafka.clients.NetworkClient)
[2025-09-12 16:19:17,066] WARN [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Bootstrap broker localhost:9092 (id: -1 rack: null) disconnected (or
g.apache.kafka.clients.NetworkClient)
[2025-09-12 16:19:17,144] INFO [MetadataLoader id=1005] initializeNewPublishers: the loader is still catching up because we still don't know the high water mark yet. (org.apac
he.kafka.image.loader.MetadataLoader)
[2025-09-12 16:19:17,172] INFO [RaftManager id=1005] Fetching snapshot OffsetAndEpoch(offset=357460, epoch=54) from Fetch response from leader 0 (org.apache.kafka.raft.KafkaRa
ftClient)第二次启动：
[2025-09-12 16:25:38,779] INFO [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Node -1 disconnected. (org.apache.kafka.clients.NetworkClient)
[2025-09-12 16:25:38,779] WARN [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Connection to node -1 (localhost/127.0.0.1:9092) could not be establ
ished. Node may not be available. (org.apache.kafka.clients.NetworkClient)
[2025-09-12 16:25:38,780] WARN [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Bootstrap broker localhost:9092 (id: -1 rack: null) disconnected (or
g.apache.kafka.clients.NetworkClient)
[2025-09-12 16:25:38,822] INFO [ObjectStorage-0-5] List objects finished, count: 0, cost: 39ms (com.automq.stream.s3.operator.AbstractObjectStorage)
[2025-09-12 16:25:38,839] INFO Reset S3 WAL (com.automq.stream.s3.wal.impl.object.ObjectWALService)
[2025-09-12 16:25:38,839] INFO Shutdown S3 WAL. (com.automq.stream.s3.wal.impl.object.ObjectWALService)
[2025-09-12 16:25:38,841] INFO S3 WAL record accumulator is closed. (com.automq.stream.s3.wal.impl.object.RecordAccumulator)
[2025-09-12 16:25:38,852] INFO Acquire permission for node: 1005, epoch: 1757665537321, failover: false (com.automq.stream.s3.wal.impl.object.ObjectReservationService)
[2025-09-12 16:25:38,918] INFO Start S3 WAL. (com.automq.stream.s3.wal.impl.object.ObjectWALService)
[2025-09-12 16:25:39,036] INFO [BrokerLifecycleManager id=1005] The broker is in RECOVERY. (kafka.server.BrokerLifecycleManager)
[2025-09-12 16:25:39,186] INFO [ObjectStorage-0-7] List objects finished, count: 0, cost: 13ms (com.automq.stream.s3.operator.AbstractObjectStorage)
[2025-09-12 16:25:39,186] INFO Register nodeId=1005 nodeEpoch=1757665537321 with new WAL configs: 0@s3://shareit-tmp-sg?region=ap-southeast-3&endpoint=https://obs.ap-southeast
-3.myhuaweicloud.com (kafka.log.stream.s3.wal.BootstrapWalV1)
[2025-09-12 16:25:39,196] INFO Reset S3 WAL (com.automq.stream.s3.wal.impl.object.ObjectWALService)
[2025-09-12 16:25:39,197] INFO S3Storage start completed (com.automq.stream.s3.S3Storage)
[2025-09-12 16:25:39,197] INFO [CompactionManager id=1005] Next Compaction started in 600000 ms (com.automq.stream.s3.compact.CompactionManager)
[2025-09-12 16:25:39,198] INFO S3Client started (kafka.log.stream.s3.DefaultS3Client)
[2025-09-12 16:25:39,206] INFO [BrokerServer id=1005] Waiting for the broker to be unfenced (kafka.server.BrokerServer)
[2025-09-12 16:25:39,214] INFO [BrokerLifecycleManager id=1005] The broker has been unfenced. Transitioning from RECOVERY to RUNNING. (kafka.server.BrokerLifecycleManager)
[2025-09-12 16:25:39,214] INFO [BrokerServer id=1005] Finished waiting for the broker to be unfenced (kafka.server.BrokerServer)
[2025-09-12 16:25:39,219] INFO authorizerStart completed for endpoint PLAINTEXT. Endpoint is now READY. (org.apache.kafka.server.network.EndpointReadyFutures)
[2025-09-12 16:25:39,219] INFO [SocketServer listenerType=BROKER, nodeId=1005] Enabling request processing. (kafka.network.SocketServer)
[2025-09-12 16:25:39,222] INFO Awaiting socket connections on 0.0.0.0:9092. (kafka.network.DataPlaneAcceptor)
[2025-09-12 16:25:39,225] INFO [BrokerServer id=1005] Waiting for all of the authorizer futures to be completed (kafka.server.BrokerServer)
[2025-09-12 16:25:39,225] INFO [BrokerServer id=1005] Finished waiting for all of the authorizer futures to be completed (kafka.server.BrokerServer)
[2025-09-12 16:25:39,226] INFO [BrokerServer id=1005] Waiting for all of the SocketServer Acceptors to be started (kafka.server.BrokerServer)
[2025-09-12 16:25:39,226] INFO [BrokerServer id=1005] Finished waiting for all of the SocketServer Acceptors to be started (kafka.server.BrokerServer)
[2025-09-12 16:25:39,226] INFO [BrokerServer id=1005] Transition from STARTING to STARTED (kafka.server.BrokerServer)
[2025-09-12 16:25:39,226] INFO Kafka version: 3.9.0 (org.apache.kafka.common.utils.AppInfoParser)
[2025-09-12 16:25:39,226] INFO Kafka commitId: 6636f8791a0d7a66 (org.apache.kafka.common.utils.AppInfoParser)
[2025-09-12 16:25:39,226] INFO Kafka startTimeMs: 1757665539226 (org.apache.kafka.common.utils.AppInfoParser)
[2025-09-12 16:25:39,227] INFO [KafkaRaftServer nodeId=1005] Kafka Server started (kafka.server.KafkaRaftServer)
[2025-09-12 16:25:39,241] INFO [MetadataLoader id=1005] InitializeNewPublishers: initializing failover-listener with a snapshot at offset 362422 (org.apache.kafka.image.loader
.MetadataLoader)
[2025-09-12 16:25:39,814] INFO [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] Cluster ID: sis-automq-test (org.apache.kafka.clients.Metadata)
[2025-09-12 16:25:39,925] INFO [Producer clientId=__automq_client_auto_balancer_metrics_reporter_producer] ProducerId set to 4001 with epoch 0 (org.apache.kafka.clients.produc
er.internals.TransactionManager)