= .truststore. =.module.Krb5LoginModule required useKeyTab=true storeKey=true serviceName="kafka" keyTab="/etc/security/keytabs/_ksqldb_uat3.keytab" Enable snappy compression for the Kafka producers # The set of Kafka brokers to bootstrap Kafka cluster information from:īrvers=broker1:9092.broker2:9092 # Automatically create a stream within KSQL for the processing log: # Automatically create the processing log topic if it does not already exist: Here's our ksql-server.properties: listeners=/etc/security/truststore/ INFO Error sending fetch request (sessionId=1228760502, epoch=1) to node 4Īnd we can also see the following log in our /var/log/messages: "Feb 24 08:10:46 server01 ksql-server-start: 8:10:46 AM io.įeb 24 08:10:46 server01 ksql-server-start: WARNING: Thread Thread=Thread has been blocked for 4552 ms, time limit is 2000 msįeb 24 08:10:46 server01 ksql-server-start: 8:10:46 AM io.įeb 24 08:10:46 server01 ksql-server-start: WARNING: Thread Thread=Thread has been blocked for 4552 ms, time limit is 2000 ms" INFO Cancelled in-flight FETCH request with correlation id 4703575 due to node 6 being disconnected (elapsed time since creation: 33666ms, elapsed time since send: 33666ms, request timeout: 30000ms) (.NetworkClient:341) This setting is distinct from the ksql.streams. INFO Disconnecting from node 6 due to request timeout. Sets the compression type used by Kafka producers, like the INSERT VALUES statement. INFO Cancelled in-flight FETCH request with correlation id 4703574 due to node 4 being disconnected (elapsed time since creation: 33672ms, elapsed time since send: 33672ms, request timeout: 30000ms) (.NetworkClient:341) INFO Disconnecting from node 4 due to request timeout.
INFO Error sending fetch request (sessionId=1099103137, epoch=1) to node 6: (.FetchSessionHandler:602) We see the following error messages in the client logs repeatedly: INFO Error sending fetch request (sessionId=1589599502, epoch=1) to node 3: (.FetchSessionHandler:602) After we restart it, it runs for about a week and then the problem occurs again.
We consume KAFKA messages in microbatches of 2 seconds.
The server status becomes "unknown," and we cannot perform queries. Below is a short Scala program and SBT build script to generate a Spark application to submit to YARN that will run on the HDP Sandbox with Spark 1.6. After about a week, the KSQLDB randomly stops working, and we are unable to use the KSQL CLI. We are running a KSQLDB with two nodes, which connects to a 16-node Kafka cluster.
0 kommentar(er)
