在分布式模式下启动Kafka连接时请求超时消息

Question

我已经下载了合流版本5.4。

和我的连接器在独立模式下运行良好。

wget /usr/local http://packages.confluent.io/archive/5.4/confluent-5.4.0-2.11.tar.gz -P ~/Downloads/
tar -zxvf ~/Downloads/confluent-5.4.0-2.11.tar.gz -C ~/Downloads/
sudo mv ~/Downloads/confluent-5.4.0 /usr/local/

然后我尝试在分布式模式下运行Kafak，所以我修改了我的connect分布。

bootstrap.servers=b-***.eu-west-1.amazonaws.com:9092,b-***.eu-west-1.amazonaws.com:9092,b-***.eu-west-1.amazonaws.com:9092
group.id=connect-cluster
key.converter=org.apache.kafka.connect.storage.StringConverter
value.converter=org.apache.kafka.connect.storage.StringConverter

offset.storage.topic=connect-offsets
offset.storage.replication.factor=1
config.storage.topic=connect-configs
config.storage.replication.factor=1
status.storage.topic=connect-status
status.storage.replication.factor=1
offset.flush.interval.ms=10000
plugin.path=/usr/local/confluent/share/java

然后启动连接器，如下所示

/usr/local/confluent/bin/connect-distributed /usr/local/confluent/etc/kafka/connect-distributed.properties

看来它成功地启动了2020-02-02 05:22:33,860 INFO加入了小组并得到了分配：

Assignment{error=0, leader='connect-1-c99d50a9-faf0-4b15-8a3d-3add55b7e206', leaderUrl='http://10.97.49.217:8083/', offset=-1, connectorIds=[], taskIds=[]} (org.apache.kafka.connect.runtime.distributed.DistributedHerder:1216)
[2020-02-02 05:22:33,861] INFO Starting connectors and tasks using config offset -1 (org.apache.kafka.connect.runtime.distributed.DistributedHerder:850)
[2020-02-02 05:22:33,861] INFO Finished starting connectors and tasks (org.apache.kafka.connect.runtime.distributed.DistributedHerder:860)

最后，我使用curl命令来创建我的接收器连接器。

curl  -X POST -H "Content-Type: application/json" --data '{"name":"elastic-search-sink-audit","config":{"connector.class":"io.confluent.connect.elasticsearch.ElasticsearchSinkConnector","tasks.max":"2","topics":"fsp-AUDIT_EVENT_DEMO","key.ignore":"true","connection.url":"https://**.amazonaws.com","type.name":"kafka-connect","name":"elastic-search-sink-audit","errors.tolerance":"all","errors.deadletterqueue.topic.name":"fsp-dlq-audit-event"}}' http://localhost:8083/connectors | jq

使用EC2 IP地址

curl  -X POST -H "Content-Type: application/json" --data '{"name":"elastic-search-sink-audit-distributed","config":{"connector.class":"io.confluent.connect.elasticsearch.ElasticsearchSinkConnector","tasks.max":"2","topics":"audit_event_distributed","key.ignore":"true","connection.url":"https://***.us-east-1.es.amazonaws.com","type.name":"kafka-connect","name":"elastic-search-sink-audit-distributed","errors.tolerance":"all","errors.deadletterqueue.topic.name":"dlq_distributed"}}' http://10.67.39.217:8083/connectors | jq

在1分钟后运行此命令后，我将得到以下响应

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   498    0    48  100   450      0      4  0:01:52  0:01:30  0:00:22    10
{
  "error_code": 500,
  "message": "Request timed out"
}

然后我继续得到它，但这只是警告，因为我有足够的副本在我的主题。

NOT_ENOUGH_REPLICAS (org.apache.kafka.clients.producer.internals.Sender:598)
^C[2020-02-03 08:05:54,890] WARN [Producer clientId=producer-3] Got error produce response with correlation id 7185 on topic-partition connect-configs-0, retrying (2147476495 attempts left). Error: NOT_ENOUGH_REPLICAS (org.apache.kafka.clients.producer.internals.Sender:598)
^C[2020-02-03 08:05:54,991] WARN [Producer clientId=producer-3] Got error produce response with correlation id 7186 on topic-partition connect-configs-0, retrying (2147476494 attempts left). Error: NOT_ENOUGH_REPLICAS (org.apache.kafka.clients.producer.internals.Sender:598)

当我描述我的主题时

Topic:fsp-AUDIT_EVENT_DEMO      PartitionCount:1        ReplicationFactor:3     Configs:
        Topic: fsp-AUDIT_EVENT_DEMO     Partition: 0    Leader: 1       Replicas: 1,6,2 Isr: 1,6,2

所以我有足够的复制品来做我的主题。

更新

描述结果

Topic:connect-configs   PartitionCount:1        ReplicationFactor:1     Configs:cleanup.policy=compact
        Topic: connect-configs  Partition: 0    Leader: 6       Replicas: 6     Isr: 6


Topic:connect-status    PartitionCount:5        ReplicationFactor:1     Configs:cleanup.policy=compact
        Topic: connect-status   Partition: 0    Leader: 6       Replicas: 6     Isr: 6
        Topic: connect-status   Partition: 1    Leader: 1       Replicas: 1     Isr: 1
        Topic: connect-status   Partition: 2    Leader: 2       Replicas: 2     Isr: 2
        Topic: connect-status   Partition: 3    Leader: 4       Replicas: 4     Isr: 4
        Topic: connect-status   Partition: 4    Leader: 5       Replicas: 5     Isr: 5


Topic:connect-offsets   PartitionCount:25       ReplicationFactor:1     Configs:cleanup.policy=compact

请帮我解决这个问题。

Answer 1

这三个主题都是由卡夫卡连接-补偿连接-信任连接状态创建的。

他们确实是

offset.storage.topic=connect-offsets
offset.storage.replication.factor=1

config.storage.topic=connect-configs
config.storage.replication.factor=1

status.storage.topic=connect-status
status.storage.replication.factor=1

，你建议我改变这个吗？

不是名字，而是复制因素，没错。1副本意味着如果任何一个代理被关闭，那么主题基本上是脱机的。

您还在哪里寻找代理id？

来自主题描述的第3-5栏。