Question

我正在开发一个使用Kafka代理和Schema Registry的GCP Cloud Dataflow作业。我们的Kafka经纪人和架构注册表需要TLS客户端证书。我在部署时面临与Schema Registry的连接问题。任何建议都受到欢迎。

这是我为Dataflow工作所做的。我为TLS配置创建了使用者属性。

props.put("security.protocol", "SSL");
props.put("ssl.truststore.password", "aaa");
props.put("ssl.keystore.password", "bbb");
props.put("ssl.key.password", "ccc"));
props.put("schema.registry.url", "https://host:port")
props.put("specific.avro.reader", true);

并通过updateConsumerProperties更新使用者属性。

Pipeline p = Pipeline.create(options)
...
.updateConsumerProperties(properties)
...

正如这个stackoverflow答案所建议的，我还将下载keyStore和trustStore到本地目录，并在ConsumerFactory中的ConsumerProperties上指定trustStore / keyStore位置。

Truststore and Google Cloud Dataflow

Pipeline p = Pipeline.create(options)
 ...
 .withConsumerFactoryFn(new MyConsumerFactory(...))
 ...

在ConsumerFactory中：

public Consumer<byte[], byte[]> apply(Map<String, Object> config)  {
  // download keyStore and trustStore from GCS bucket 
  config.put("ssl.truststore.location", (Object)localTrustStoreFilePath)
  config.put("ssl.keystore.location", (Object)localKeyStoreFilePath)
  new KafkaConsumer<byte[], byte[]>(config);
}

使用此代码，我成功部署了，但是Dataflow作业遇到了TLS服务器证书验证错误。

Caused by: sun.security.validator.ValidatorException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target
        sun.security.validator.PKIXValidator.doBuild(PKIXValidator.java:387)
        sun.security.validator.PKIXValidator.engineValidate(PKIXValidator.java:292)
        sun.security.validator.Validator.validate(Validator.java:260)
        sun.net.www.protocol.https.HttpsClient.afterConnect(HttpsClient.java:559)
        sun.net.www.protocol.https.AbstractDelegateHttpsURLConnection.connect(AbstractDelegateHttpsURLConnection.java:185)
        sun.net.www.protocol.http.HttpURLConnection.getInputStream0(HttpURLConnection.java:1513)
        sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1441)
        java.net.HttpURLConnection.getResponseCode(HttpURLConnection.java:480)
        sun.net.www.protocol.https.HttpsURLConnectionImpl.getResponseCode(HttpsURLConnectionImpl.java:338)
        io.confluent.kafka.schemaregistry.client.rest.RestService.sendHttpRequest(RestService.java:208)
        io.confluent.kafka.schemaregistry.client.rest.RestService.httpRequest(RestService.java:252)
        io.confluent.kafka.schemaregistry.client.rest.RestService.getId(RestService.java:482)
        io.confluent.kafka.schemaregistry.client.rest.RestService.getId(RestService.java:475)
        io.confluent.kafka.schemaregistry.client.CachedSchemaRegistryClient.getSchemaByIdFromRegistry(CachedSchemaRegistryClient.java:151)
        io.confluent.kafka.schemaregistry.client.CachedSchemaRegistryClient.getBySubjectAndId(CachedSchemaRegistryClient.java:230)
        io.confluent.kafka.schemaregistry.client.CachedSchemaRegistryClient.getById(CachedSchemaRegistryClient.java:209)
        io.confluent.kafka.serializers.AbstractKafkaAvroDeserializer.deserialize(AbstractKafkaAvroDeserializer.java:116)
        io.confluent.kafka.serializers.AbstractKafkaAvroDeserializer.deserialize(AbstractKafkaAvroDeserializer.java:88)
        org.fastretailing.rfid.store.siv.EPCTransactionKafkaAvroDeserializer.deserialize(EPCTransactionKafkaAvroDeserializer.scala:14)
        org.fastretailing.rfid.store.siv.EPCTransactionKafkaAvroDeserializer.deserialize(EPCTransactionKafkaAvroDeserializer.scala:7)
        org.apache.beam.sdk.io.kafka.KafkaUnboundedReader.advance(KafkaUnboundedReader.java:234)
        org.apache.beam.sdk.io.kafka.KafkaUnboundedReader.start(KafkaUnboundedReader.java:176)
        org.apache.beam.runners.dataflow.worker.WorkerCustomSources$UnboundedReaderIterator.start(WorkerCustomSources.java:779)
        org.apache.beam.runners.dataflow.worker.util.common.worker.ReadOperation$SynchronizedReaderIterator.start(ReadOperation.java:361)
        org.apache.beam.runners.dataflow.worker.util.common.worker.ReadOperation.runReadLoop(ReadOperation.java:194)
        org.apache.beam.runners.dataflow.worker.util.common.worker.ReadOperation.start(ReadOperation.java:159)
        org.apache.beam.runners.dataflow.worker.util.common.worker.MapTaskExecutor.execute(MapTaskExecutor.java:76)
        org.apache.beam.runners.dataflow.worker.StreamingDataflowWorker.process(StreamingDataflowWorker.java:1228)
        org.apache.beam.runners.dataflow.worker.StreamingDataflowWorker.access$1000(StreamingDataflowWorker.java:143)
        org.apache.beam.runners.dataflow.worker.StreamingDataflowWorker$6.run(StreamingDataflowWorker.java:967)
        java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
        java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
        java.lang.Thread.run(Thread.java:745)

然后，我发现Schema Registry客户端从系统属性加载TLS配置。 https://github.com/confluentinc/schema-registry/issues/943

我以相同的配置测试了Kafka Consumer，并确认它可以正常工作。

props.put("schema.registry.url", "https://host:port")
props.put("specific.avro.reader", true);
props.put("ssl.truststore.location", System.getProperty("javax.net.ssl.trustStore"));
props.put("ssl.truststore.password", System.getProperty("javax.net.ssl.keyStore"));
props.put("ssl.keystore.location", System.getProperty("javax.net.ssl.keyStore"));
props.put("ssl.keystore.password", System.getProperty("javax.net.ssl.keyStorePassword"));
props.put("ssl.key.password", System.getProperty("javax.net.ssl.key.password"));

接下来，我应用了相同的方法，这意味着将相同的TLS配置应用于数据流作业代码的系统属性和使用者属性。

我在执行应用程序时通过系统属性指定了密码。

-Djavax.net.ssl.keyStorePassword=aaa \
-Djavax.net.ssl.key.password=bbb \
-Djavax.net.ssl.trustStorePassword=ccc \

注意：由于这些文件已下载到本地临时目录，因此我在Consumer Factory中为trustStore和keyStore位置设置了系统属性。

config.put("ssl.truststore.location", (Object)localTrustStoreFilePath)
config.put("ssl.keystore.location", (Object)localKeyStoreFilePath)
System.setProperty("javax.net.ssl.trustStore", localTrustStoreFilePath)
System.setProperty("javax.net.ssl.keyStore", localKeyStoreFilePath)

但部署失败并出现超时错误。

Exception in thread "main" java.lang.RuntimeException: Failed to construct instance from factory method DataflowRunner#fromOptions(interface org.apache.beam.sdk.options.PipelineOptions)
        at org.apache.beam.sdk.util.InstanceBuilder.buildFromMethod(InstanceBuilder.java:224)
...
Caused by: java.lang.reflect.InvocationTargetException
        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
...
Caused by: java.lang.IllegalArgumentException: DataflowRunner requires gcpTempLocation, but failed to retrieve a value from PipelineOptions
        at org.apache.beam.runners.dataflow.DataflowRunner.fromOptions(DataflowRunner.java:246)
Caused by: java.lang.IllegalArgumentException: Error constructing default value for gcpTempLocation: tempLocation is not a valid GCS path, gs://dev-k8s-rfid-store-dataflow/rfid-store-siv-epc-transactions-to-bq/tmp. 
        at org.apache.beam.sdk.extensions.gcp.options.GcpOptions$GcpTempLocationFactory.create(GcpOptions.java:255)
...
Caused by: java.lang.RuntimeException: Unable to verify that GCS bucket gs://dev-k8s-rfid-store-dataflow exists.
        at org.apache.beam.sdk.extensions.gcp.storage.GcsPathValidator.verifyPathIsAccessible(GcsPathValidator.java:86)
...
Caused by: java.io.IOException: Error getting access token for service account: java.security.NoSuchAlgorithmException: Error constructing implementation (algorithm: Default, provider: SunJSSE, class: sun.security.ssl.SSLContextImpl$DefaultSSLContext)
        at com.google.auth.oauth2.ServiceAccountCredentials.refreshAccessToken(ServiceAccountCredentials.java:401)
...
Caused by: java.net.SocketException: java.security.NoSuchAlgorithmException: Error constructing implementation (algorithm: Default, provider: SunJSSE, class: sun.security.ssl.SSLContextImpl$DefaultSSLContext)
        at javax.net.ssl.DefaultSSLSocketFactory.throwException(SSLSocketFactory.java:248)
...
Caused by: java.security.NoSuchAlgorithmException: Error constructing implementation (algorithm: Default, provider: SunJSSE, class: sun.security.ssl.SSLContextImpl$DefaultSSLContext)
        at java.security.Provider$Service.newInstance(Provider.java:1617)
...
Caused by: java.io.IOException: Keystore was tampered with, or password was incorrect
    at sun.security.provider.JavaKeyStore.engineLoad(JavaKeyStore.java:780)
Caused by: java.security.UnrecoverableKeyException: Password verification failed
    at sun.security.provider.JavaKeyStore.engineLoad(JavaKeyStore.java:778)

我想念什么吗？

Answer 1

在ConsumerFactoryFn中，您需要将证书从某个位置（例如GCS）复制到计算机上的本地文件路径。

在FileNotFoundException中，用户编写的ConsumerFnFactory具有以下代码段，该代码段可从GCS中获取信任库：

            Storage storage = StorageOptions.newBuilder()
                    .setProjectId("prj-id-of-your-bucket")
                    .setCredentials(GoogleCredentials.getApplicationDefault())
                    .build()
                    .getService();
            Blob blob = storage.get("your-bucket-name", "pth.to.your.kafka.client.truststore.jks");
            ReadChannel readChannel = blob.reader();
            FileOutputStream fileOuputStream;
            fileOuputStream = new FileOutputStream("/tmp/kafka.client.truststore.jks"); //path where the jks file will be stored
            fileOuputStream.getChannel().transferFrom(readChannel, 0, Long.MAX_VALUE);
            fileOuputStream.close();
            File f = new File("/tmp/kafka.client.truststore.jks"); //assuring the store file exists
            if (f.exists())
            {
                LOG.debug("key exists");

            }
            else
            {
                LOG.error("key does not exist");

            }

您需要执行类似的操作（不一定是GCS，但确实需要从在Google Cloud Dataflow上执行管道的所有VM进行访问）。

Answer 2

我收到了GCP支持的回复。看来Apache Beam不支持Schema Registry。

你好，   Dataflow专家已经联系了我。我现在将揭露他们告诉我的内容。

您的问题的答案是否定的，Apache Beam不支持Schema Registry。   但是，他们告诉我您可以实现对Schema Registry的调用   因为Beam仅消耗原始消息，因此用户有责任自己做   他们需要什么数据。

这是基于我们对您要向Kafka发布消息的情况的理解，   并让DF使用这些消息，并根据注册表中的架构对其进行解析。

希望这些信息对您有用，请告诉我是否可以提供进一步的帮助。

但是Dataflow作业仍可以接收Avro格式的二进制消息。因此，您可以在内部内部调用Schema Registry REST API，如下所示。 https://stackoverflow.com/a/55917157

当架构注册表需要TLS客户端身份验证时，无法从数据流作业连接到架构注册表

2 个答案: