kafka连接:多个DB2JDBC源连接器失败

tkclm6bt  于 2021-06-07  发布在  Kafka
关注(0)|答案(1)|浏览(481)

我正在尝试在本地docker容器中使用kafkaconnect(使用官方合流映像),以便将db2数据推送到openshift(aws)上的kafka集群。我将合流jdbc连接器与db2jdbcjar一起使用。我有不同的连接器配置,因为我使用smt和“transforms.createkey”(创建我的键),并且我的表中的键列有不同的名称。
以下是我的步骤:
为kafka connect创建配置、偏移和状态的主题
启动/创建kafka connect容器(带env vars,见下文)
通过对我的connect容器的post调用创建第一个jdbc连接器(请参见下面的配置)
到目前为止,一切正常,我可以看到我的数据被推送到集群。但是,当我通过post调用添加第二个jdbc连接器时,第一个连接器停止向集群推送数据,而第二个连接器开始并继续加载和推送数据。在很短的一段时间内,两个连接器似乎都会将数据推送到集群中,但我假设这可能是来自连接器1的数据,仍然被刷新。问题是a)即使是跟踪日志也不会显示有意义的错误(至少对我来说)和b)在尝试之间显示的错误不同(我总是删除所有主题和容器)。
我假设这不是一个bug,而是需要适当设置的配置的组合和/或我对一些基本的kafka connect核心功能缺乏了解。我已经尝试添加和更改各种配置,但不幸的是,到目前为止没有任何结果。我试了很多次,但运气不好。我已经附上了我最近两次尝试的日志以及配置。
有没有人知道我可以修改哪个配置,或者为了解决这个问题需要研究什么?感谢您的帮助-谢谢!

Kafka: 2.0.0
Docker image: confluentinc/cp-kafka-connect:5.0.0
DB2: 10.5
JDBC Jar: db2jcc4.jar with version 4.19.76

第一次尝试:

[2018-12-17 13:09:15,683] ERROR Invalid call to OffsetStorageWriter flush() while already flushing, the framework should not allow this (org.apache.kafka.connect.storage.OffsetStorageWriter)
[2018-12-17 13:09:15,684] ERROR WorkerSourceTask{id=db2-jdbc-source-0} Task threw an uncaught and unrecoverable exception (org.apache.kafka.connect.runtime.WorkerTask)
org.apache.kafka.connect.errors.ConnectException: OffsetStorageWriter is already flushing
    at org.apache.kafka.connect.storage.OffsetStorageWriter.beginFlush(OffsetStorageWriter.java:110)
    at org.apache.kafka.connect.runtime.WorkerSourceTask.commitOffsets(WorkerSourceTask.java:409)
    at org.apache.kafka.connect.runtime.WorkerSourceTask.execute(WorkerSourceTask.java:238)
    at org.apache.kafka.connect.runtime.WorkerTask.doRun(WorkerTask.java:175)
    at org.apache.kafka.connect.runtime.WorkerTask.run(WorkerTask.java:219)
    at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
    at java.util.concurrent.FutureTask.run(FutureTask.java:266)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at java.lang.Thread.run(Thread.java:748)
[2018-12-17 13:09:15,686] ERROR WorkerSourceTask{id=db2-jdbc-source-0} Task is being killed and will not recover until manually restarted (org.apache.kafka.connect.runtime.WorkerTask)
[2018-12-17 13:09:15,686] INFO [Producer clientId=producer-4] Closing the Kafka producer with timeoutMillis = 30000 ms. (org.apache.kafka.clients.producer.KafkaProducer)
[2018-12-17 13:09:20,682] ERROR Graceful stop of task db2-jdbc-source-0 failed. (org.apache.kafka.connect.runtime.Worker)
[2018-12-17 13:09:20,682] INFO Finished stopping tasks in preparation for rebalance (org.apache.kafka.connect.runtime.distributed.DistributedHerder)

第二次尝试:

[2018-12-17 14:01:31,658] INFO Stopping task db2-jdbc-source-0 (org.apache.kafka.connect.runtime.Worker)
[2018-12-17 14:01:31,689] INFO Stopped connector db2-jdbc-source (org.apache.kafka.connect.runtime.Worker)
[2018-12-17 14:01:31,784] INFO WorkerSourceTask{id=db2-jdbc-source-0} Committing offsets (org.apache.kafka.connect.runtime.WorkerSourceTask)
[2018-12-17 14:01:31,784] INFO WorkerSourceTask{id=db2-jdbc-source-0} flushing 20450 outstanding messages for offset commit (org.apache.kafka.connect.runtime.WorkerSourceTask)
[2018-12-17 14:01:36,733] ERROR Graceful stop of task db2-jdbc-source-0 failed. (org.apache.kafka.connect.runtime.Worker)
[2018-12-17 14:01:36,733] INFO Finished stopping tasks in preparation for rebalance (org.apache.kafka.connect.runtime.distributed.DistributedHerder)

Kafka集群中每秒传入消息的屏幕截图
kafka connect docker环境变量:

-e CONNECT_BOOTSTRAP_SERVERS=my_kafka_cluster:443 \
  -e CONNECT_PRODUCER_BOOTSTRAP_SERVERS="my_kafka_cluster:443" \
  -e CONNECT_REST_ADVERTISED_HOST_NAME="kafka-connect" \
  -e CONNECT_REST_PORT=8083 \
  -e CONNECT_GROUP_ID="kafka-connect-group" \
  -e CONNECT_CONFIG_STORAGE_REPLICATION_FACTOR=3 \
  -e CONNECT_CONFIG_STORAGE_TOPIC="kafka-connect-config" \
  -e CONNECT_OFFSET_STORAGE_REPLICATION_FACTOR=3 \
  -e CONNECT_OFFSET_STORAGE_TOPIC="kafka-connect-offset" \
  -e CONNECT_OFFSET_FLUSH_INTERVAL_MS=15000 \
  -e CONNECT_OFFSET_FLUSH_TIMEOUT_MS=60000 \
  -e CONNECT_STATUS_STORAGE_REPLICATION_FACTOR=3 \
  -e CONNECT_STATUS_STORAGE_TOPIC="kafka-connect-status" \
  -e CONNECT_KEY_CONVERTER="io.confluent.connect.avro.AvroConverter" \
  -e CONNECT_KEY_CONVERTER_SCHEMA_REGISTRY_URL=http://url_to_schemaregistry \
  -e CONNECT_VALUE_CONVERTER="io.confluent.connect.avro.AvroConverter" \
  -e CONNECT_VALUE_CONVERTER_SCHEMA_REGISTRY_URL=http://url_to_schemaregistry \
  -e CONNECT_INTERNAL_KEY_CONVERTER="org.apache.kafka.connect.json.JsonConverter" \
  -e CONNECT_INTERNAL_KEY_CONVERTER_SCHEMAS_ENABLE="false" \
  -e CONNECT_INTERNAL_VALUE_CONVERTER="org.apache.kafka.connect.json.JsonConverter" \
  -e CONNECT_INTERNAL_VALUE_CONVERTER_SCHEMAS_ENABLE="false" \
  -e CONNECT_PLUGIN_PATH=/usr/share/java \
  -e CONNECT_PRODUCER_BUFFER_MEMORY="8388608" \
  -e CONNECT_SECURITY_PROTOCOL="SSL" \
  -e CONNECT_PRODUCER_SECURITY_PROTOCOL="SSL" \
  -e CONNECT_SSL_TRUSTSTORE_LOCATION="/usr/share/kafka.client.truststore.jks" \
  -e CONNECT_PRODUCER_SSL_TRUSTSTORE_LOCATION="/usr/share/kafka.client.truststore.jks" \
  -e CONNECT_SSL_TRUSTSTORE_PASSWORD="my_ts_pw" \
  -e CONNECT_PRODUCER_SSL_TRUSTSTORE_PASSWORD="my_ts_pw" \
  -e CONNECT_LOG4J_LOGGERS=org.apache.kafka.connect.runtime.rest=WARN,org.reflections=ERROR \
  -e CONNECT_LOG4J_ROOT_LOGLEVEL=INFO \
  -e HOSTNAME=kafka-connect \

jdbc连接器(只有表和键列不同):

{
    "name": "db2-jdbc-source",
    "config": 
    {
        "mode":"timestamp",
        "debug":"true",
        "batch.max.rows":"50",
        "poll.interval.ms":"10000",
        "timestamp.delay.interval.ms":"60000",
        "timestamp.column.name":"IBMSNAP_LOGMARKER",
        "connector.class":"io.confluent.connect.jdbc.JdbcSourceConnector" ,
        "connection.url":"jdbc:db2://myip:myport/mydb:currentSchema=myschema;",
        "connection.password":"mypw",
        "connection.user":"myuser",
        "connection.backoff.ms":"60000",
        "dialect.name": "Db2DatabaseDialect",
        "table.types": "TABLE",
        "table.poll.interval.ms":"60000",
        "table.whitelist":"MYTABLE1",
        "tasks.max":"1",
        "topic.prefix":"db2_",
        "key.converter":"io.confluent.connect.avro.AvroConverter",
        "key.converter.schema.registry.url":"http://url_to_schemaregistry",
        "value.converter":"io.confluent.connect.avro.AvroConverter",
        "value.converter.schema.registry.url":"http://url_to_schemaregistry",
        "transforms":"createKey",
        "transforms.createKey.type":"org.apache.kafka.connect.transforms.ValueToKey",
        "transforms.createKey.fields":"MYKEY1"
    }
}
az31mfrm

az31mfrm1#

我最终解决了问题:我在timestamp模式下使用jdbc连接器,而不是timestamp+递增,因为我不能(总是)指定递增列。我知道这可能会导致问题,当有多个条目具有相同的时间戳时,connect无法知道哪些条目已经被读取。
我的大部分数据行都有相同的时间戳。当我添加第二个连接器时,存储了第一个连接器的当前时间戳,connect开始重新平衡,因此丢失了已经读取了该stimestamp的行的信息。当连接器启动并再次运行时,第一个连接器继续使用“下一个时间戳”,因此只加载最新的行(这只是一小部分)。
我的错误是假设,在这样的情况下,第一个连接器将重新使用前一个时间戳,而不是继续使用“下一个时间戳”。对我来说,宁可冒重复的风险,也不要有可能丢失数据。

相关问题