[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14202173#comment-14202173 ] schandr commented on KAFKA-1738: The patch works and thanks for the help!!!. Will this be patch be included in 0.8.2 release? If yes, do you have a date in mind? Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Assignee: Jun Rao Priority: Blocker Fix For: 0.8.2 Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt, kafka-1738.patch We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092.
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14200510#comment-14200510 ] Jun Rao commented on KAFKA-1738: You actually found a real bug, thanks! We exposed an existing problem after adding the ability to kill idle connections in KAFKA-1282. The default max idle time happens to be 10 minutes. That's why you only see the issue if the topics are created more than 10 mins apart. I will attach a patch soon. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] -
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14200512#comment-14200512 ] Jun Rao commented on KAFKA-1738: Created reviewboard https://reviews.apache.org/r/27690/diff/ against branch origin/trunk Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt, kafka-1738.patch We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14200522#comment-14200522 ] Jun Rao commented on KAFKA-1738: Sri, Pradeep, Do you think you can try the patch and see if this fixes your issue? Also marking this as an 0.8.2 blocker. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Assignee: Jun Rao Priority: Blocker Fix For: 0.8.2 Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt, kafka-1738.patch We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092.
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14200603#comment-14200603 ] schandr commented on KAFKA-1738: GreatAnd thank you for the Patch. How should we apply this patch? Are there any instructions on how to apply the patch. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Assignee: Jun Rao Priority: Blocker Fix For: 0.8.2 Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt, kafka-1738.patch We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092.
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14200708#comment-14200708 ] Neha Narkhede commented on KAFKA-1738: -- Good catch. Thanks for following up on this, Jun. Reviewed the patch, looks good. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Assignee: Jun Rao Priority: Blocker Fix For: 0.8.2 Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt, kafka-1738.patch We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker.
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14200748#comment-14200748 ] schandr commented on KAFKA-1738: Will apply the patch against the 0.8.2-beta and post the update. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Assignee: Jun Rao Priority: Blocker Fix For: 0.8.2 Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt, kafka-1738.patch We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread)
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14198592#comment-14198592 ] Jun Rao commented on KAFKA-1738: The log shows that the log dir was created for topic_A. Were you looking at the right directory? [2014-11-04 21:36:30,724] INFO Created log for partition [topic_A,0] in /tmp/kafka-logs with properties Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14198648#comment-14198648 ] Pradeep commented on KAFKA-1738: I suggest you to run the script once on a working setup of kafka. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at kafka.network.BlockingChannel.send(BlockingChannel.scala:97) at kafka.controller.RequestSendThread.liftedTree1$1(ControllerChannelManager.scala:132) at
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14198678#comment-14198678 ] Jun Rao commented on KAFKA-1738: Yes, I started a kafka broker and ran your script. I was able to see local logs created. ls /tmp/kafka-logs/ juntopic-0 test-0 topic3-0topic_3-0 recovery-point-offset-checkpointtopic1-0 topic_1-0 topic_4-0 replication-offset-checkpoint topic2-0 topic_2-0 topic_5-0 The error you saw typically happens when a broker is down. Can you telnet to host DMIPVM on port 9092 when the broker is up? Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14198688#comment-14198688 ] Pradeep commented on KAFKA-1738: Can you provide the configuration files? We are using the default configurations. Also, i tried telnet and the broker is up. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at kafka.network.BlockingChannel.send(BlockingChannel.scala:97) at
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14198713#comment-14198713 ] schandr commented on KAFKA-1738: Here are the additional logs for the same issue Controller.log [2014-11-05 10:31:12,441] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:localhost.localdomain,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-05 10:31:12,445] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 7 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:7;CorrelationId:8;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:localhost.localdomain,port:9092;PartitionState:[topic_30,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:7),ReplicationFactor:1),AllReplicas:0),[topic_5,0] - (LeaderAndIsrInfo:(Leader:-2,ISR:0,LeaderEpoch:0,ControllerEpoch:7),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:localhost.localdomain,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at kafka.network.BlockingChannel.send(BlockingChannel.scala:97) at kafka.controller.RequestSendThread.liftedTree1$1(ControllerChannelManager.scala:132) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:131) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-05 10:31:12,448] INFO [Controller-0-to-broker-0-send-thread], Controller 0 connected to id:0,host:localhost.localdomain,port:9092 for sending state change requests (kafka.controller.RequestSendThread) Server.log [2014-11-05 10:31:12,414] DEBUG Got notification sessionid:0x14980b08c110003 (org.apache.zookeeper.ClientCnxn) [2014-11-05 10:31:12,415] DEBUG Got WatchedEvent state:SyncConnected type:NodeChildrenChanged path:/brokers/topics for sessionid 0x14980b08c110003 (org.apache.zookeeper.ClientCnxn) [2014-11-05 10:31:12,415] DEBUG Received event: WatchedEvent state:SyncConnected type:NodeChildrenChanged path:/brokers/topics (org.I0Itec.zkclient.ZkClient) [2014-11-05 10:31:12,415] DEBUG New event: ZkEvent[Children of /brokers/topics changed sent to kafka.controller.PartitionStateMachine$TopicChangeListener@1fc5681] (org.I0Itec.zkclient.ZkEventThread) [2014-11-05 10:31:12,415] DEBUG Leaving process event (org.I0Itec.zkclient.ZkClient) [2014-11-05 10:31:12,415] DEBUG Delivering event #3 ZkEvent[Children of /brokers/topics changed sent to kafka.controller.PartitionStateMachine$TopicChangeListener@1fc5681] (org.I0Itec.zkclient.ZkEventThread) [2014-11-05 10:31:12,415] DEBUG Got ping response for sessionid: 0x14980b08c110003 after 0ms (org.apache.zookeeper.ClientCnxn) [2014-11-05 10:31:12,416] DEBUG Reading reply sessionid:0x14980b08c110003, packet:: clientPath:null serverPath:null finished:false header:: 257,3 replyHeader:: 257,13610,0 request:: '/brokers/topics,T response:: s{6,6,1415130748279,1415130748279,0,23,0,0,0,23,13610} (org.apache.zookeeper.ClientCnxn) [2014-11-05 10:31:12,417] DEBUG Reading reply sessionid:0x14980b08c110003, packet:: clientPath:null serverPath:null finished:false header:: 258,8 replyHeader:: 258,13610,0 request:: '/brokers/topics,T response:: v{'topic_23,'topic_18,'topic_22,'topic_17,'topic_25,'topic_16,'topic_24,'topic_15,'topic_14,'topic_13,'topic_12,'topic_11,'topic_19,'topic_5,'topic_7,'topic_6,'Test1,'topic_10,'topic_9,'topic_8,'topic_20,'topic_30,'topic_21} (org.apache.zookeeper.ClientCnxn) [2014-11-05 10:31:12,420] DEBUG Reading reply sessionid:0x14980b08c110003, packet:: clientPath:null serverPath:null finished:false header:: 259,4 replyHeader:: 259,13610,0 request:: '/brokers/topics/topic_30,F response:: #7b2276657273696f6e223a312c22706172746974696f6e73223a7b2230223a5b305d7d7d,s{13610,13610,1415205072414,1415205072414,0,0,0,0,36,0,13610} (org.apache.zookeeper.ClientCnxn) [2014-11-05 10:31:12,422] DEBUG Replicas assigned to topic [topic_30], partition [0] are [List(0)] (kafka.utils.ZkUtils$) [2014-11-05 10:31:12,422] DEBUG Replicas assigned to topic [topic_30], partition [0] are [List(0)] (kafka.utils.ZkUtils$) [2014-11-05 10:31:12,425] DEBUG Reading reply
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14198747#comment-14198747 ] schandr commented on KAFKA-1738: And I was able to telnet to the host:port. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at kafka.network.BlockingChannel.send(BlockingChannel.scala:97) at kafka.controller.RequestSendThread.liftedTree1$1(ControllerChannelManager.scala:132) at
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14199132#comment-14199132 ] schandr commented on KAFKA-1738: OkHere is my understanding. This might probably be a bug. 1. For any requests that the controller sends to the Broker, it uses the BlockingChannel - that's initialized with the controller.socket.timeout.ms value specified in the server.properties. 2. Once the channel is established the RequestSendThread uses this channel to send any requests such as LeaderAndIsr, UpdateMetaData without checking if the channel is still open 3. Based on the value, the socket might have timed out. The following Code in the RequestSendThread calls the connectToBroker again on catching the exception, but does not send the failed request. If the request happens to be LeaderAndIsr for the new partition it results in missing log directory creation or other errors which results in producer or consumer throwing exceptions when they try to produce or consume data from the failed topic. var isSendSuccessful = false while(isRunning.get() !isSendSuccessful) { // if a broker goes down for a long time, then at some point the controller's zookeeper listener will trigger a // removeBroker which will invoke shutdown() on this thread. At that point, we will stop retrying. try { channel.send(request) isSendSuccessful = true } catch { case e: Throwable = // if the send was not successful, reconnect to broker and resend the message error((Controller %d epoch %d failed to send request %s to broker %s. + Reconnecting to broker.).format(controllerId, controllerContext.epoch, request.toString, toBroker.toString()), e) channel.disconnect() connectToBroker(toBroker, channel) isSendSuccessful = false // backoff before retrying the connection and send Utils.swallow(Thread.sleep(300)) } } In the code below, after reconnecting to the broker, it should also resend the failed request. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)]
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14199625#comment-14199625 ] Jun Rao commented on KAFKA-1738: Hmm, that try/catch block is in a while loop. So, the resend should happen when the logic loops back again. Also, normally when a broker goes down, the controller typically notices it immediately and will remove the corresponding BlockingChannel. When a broker comes back, a new BlockingChannel will be created. It sounds like that you see this issue in some broker failure scenarios. Could you describe that a bit more? Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14199643#comment-14199643 ] schandr commented on KAFKA-1738: Thank you very much for your response. First I think the broker is not down. Because I am able to telnet to the VM where kafka is running using the host:9092. We have tried several scenarios 1. Creating a topic through Java code using TopicCommand 2. Creating a topic in a bash script. In either of the cases above, the topic creation fails randomly. When the bash script is modified to create the topic every 10 minutes, it fails consistently. In the server log, I see the request to LeaderAndISR fails, that results in partition log file not getting created. Only zookeeper and Kafka is running in the VM where the script was ran. Let me know if you need more information like server.properties file. Also I am able to see the zk nodes for the controller and broker through an eclipse zookeeper plugin Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14199662#comment-14199662 ] Jun Rao commented on KAFKA-1738: What's your value for controller.socket.timeout.ms? Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at kafka.network.BlockingChannel.send(BlockingChannel.scala:97)
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14199678#comment-14199678 ] schandr commented on KAFKA-1738: controller.socket.timeout.ms=9 controller.message.queue.size=20 auto.leader.rebalance.enable=true queued.max.requests=20 Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14199702#comment-14199702 ] schandr commented on KAFKA-1738: One more observance in the server log Here is the Trace statement for the send request. This gets logged when channel.send(request) is invoked in the RequestSendThread. The send method invokes writeCompletely method in the Send class. 0:31:12,440] TRACE 131 bytes written. (kafka.network.BoundedByteBufferSend) [2014-11-05 10:31:12,440] TRACE 131 bytes written. (kafka.network.BoundedByteBufferSend) If the request gets resend the same statement should be relogged in the server log based on channel.send(request) code in the RequestSendThread which did not get logged. Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep Attachments: 1738.zip, ServerLogForFailedTopicCreation.txt, ServerLogForFailedTopicCreation.txt, ServerLogForSuccessfulTopicCreation.txt We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14196328#comment-14196328 ] Pradeep commented on KAFKA-1738: yes.. This is happening in 8.2.0-beta version which we are currently on. I had two topic in kakfa before killing the service and i added three more topics (ATopic, BTopic and CTopic) and all of them got created successfully as you can see below. When i list it i can see the topics in kakfa. But if i see the data folder in kafka, i dont see the partition folders/files for them. Let me know if you need more details. [root@dmipvm temp]# service kafka status Kafka is running as 19396. LISTEN on tcp port= LISTEN on tcp port=53536 LISTEN on tcp port=9092 LISTEN on tcp port=48330 [root@dmipvm temp]# kill -9 19396 [root@dmipvm temp]# service kafka start Starting kafka ... STARTED. [root@dmipvm temp]# /apps/kafka/bin/kafka-topics.sh --create --topic ATopic --partitions 1 --replication-factor 1 --zookeeper localhost:2181 Created topic ATopic. [root@dmipvm temp]# /apps/kafka/bin/kafka-topics.sh --create --topic BTopic --partitions 1 --replication-factor 1 --zookeeper localhost:2181 Created topic BTopic. [root@dmipvm temp]# /apps/kafka/bin/kafka-topics.sh --create --topic CTopic --partitions 1 --replication-factor 1 --zookeeper localhost:2181 Created topic CTopic. [root@dmipvm temp]# [root@DMIPVM kafka]# ls -lrt total 16 drwxr-xr-x 2 root root 4096 Nov 4 11:32 topic_1-0 drwxr-xr-x 2 root root 4096 Nov 4 11:34 topic_2-0 -rw-r--r-- 1 root root 28 Nov 4 11:35 replication-offset-checkpoint -rw-r--r-- 1 root root 28 Nov 4 11:36 recovery-point-offset-checkpoint [root@DMIPVM kafka]# /apps/kafka/bin/kafka-topics.sh --list --zookeeper localhost:2181 ATopic BTopic CTopic topic_1 topic_2 Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14195577#comment-14195577 ] Jun Rao commented on KAFKA-1738: I can't reproduce this issue by following the steps in the description. Does this happen every time? Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. We see issues when we are creating new topics. This happens randomly and we dont know the exact reasons. We see the below logs in controller during the time of creation of topics which doesnt have the partition files. 2014-11-03 13:12:50,625] INFO [Controller 0]: New topic creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Controller 0]: New partition creation callback for [JobJTopic,0] (kafka.controller.KafkaController) [2014-11-03 13:12:50,626] INFO [Partition state machine on Controller 0]: Invoking state change to NewPartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,653] INFO [Replica state machine on controller 0]: Invoking state change to NewReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,654] INFO [Partition state machine on Controller 0]: Invoking state change to OnlinePartition for partitions [JobJTopic,0] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Live assigned replicas for partition [JobJTopic,0] are: [List(0)] (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,654] DEBUG [Partition state machine on Controller 0]: Initializing leader and isr for partition [JobJTopic,0] to (Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2) (kafka.controller.PartitionStateMachine) [2014-11-03 13:12:50,667] INFO [Replica state machine on controller 0]: Invoking state change to OnlineReplica for replicas [Topic=JobJTopic,Partition=0,Replica=0] (kafka.controller.ReplicaStateMachine) [2014-11-03 13:12:50,794] WARN [Controller-0-to-broker-0-send-thread], Controller 0 fails to send a request to broker id:0,host:DMIPVM,port:9092 (kafka.controller.RequestSendThread) java.io.EOFException: Received -1 when reading from channel, socket has likely been closed. at kafka.utils.Utils$.read(Utils.scala:381) at kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54) at kafka.network.Receive$class.readCompletely(Transmission.scala:56) at kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29) at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108) at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146) at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) [2014-11-03 13:12:50,965] ERROR [Controller-0-to-broker-0-send-thread], Controller 0 epoch 2 failed to send request Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:2;CorrelationId:43;ClientId:id_0-host_null-port_9092;AliveBrokers:id:0,host:DMIPVM,port:9092;PartitionState:[JobJTopic,0] - (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:2),ReplicationFactor:1),AllReplicas:0) to broker id:0,host:DMIPVM,port:9092. Reconnecting to broker. (kafka.controller.RequestSendThread) java.nio.channels.ClosedChannelException at kafka.network.BlockingChannel.send(BlockingChannel.scala:97) at kafka.controller.RequestSendThread.liftedTree1$1(ControllerChannelManager.scala:132)
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14190370#comment-14190370 ] Neha Narkhede commented on KAFKA-1738: -- [~pradeepbadiger] Do you also try to send data to the topic you are hoping to see the logs for? Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KAFKA-1738) Partitions for topic not created after restart from forced shutdown
[ https://issues.apache.org/jira/browse/KAFKA-1738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14191229#comment-14191229 ] Pradeep commented on KAFKA-1738: Yes. We tried sending data but it gave the exception. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic JobCTopic - No partition metadata for topic JobCTopic due to kafka.common.UnknownTopicOrPartitionException}] for topic [JobCTopic]: class kafka.common.UnknownTopicOrPartitionException Partitions for topic not created after restart from forced shutdown --- Key: KAFKA-1738 URL: https://issues.apache.org/jira/browse/KAFKA-1738 Project: Kafka Issue Type: Bug Components: core Affects Versions: 0.8.1.1, 0.8.2 Environment: Linux, 2GB RAM, 2 Core CPU Reporter: Pradeep We are using Kafka Topic APIs to create the topic. But in some cases, the topic gets created but we don't see the partition specific files and when producer/consumer tries to get the topic metadata and it fails with exception. Same happens if one tries to create using the command line. k.p.BrokerPartitionInfo - Error while fetching metadata [{TopicMetadata for topic tloader1 - No partition metadata for topic tloader1 due to kafka.common.UnknownTopicOrPartitionException}] for topic [tloader1]: class kafka.common.UnknownTopicOrPartitionException Steps to reproduce - 1. Stop kafka using kill -9 PID of Kafka 2. Start Kafka 3. Create Topic with partition and replication factor of 1. 4. Check the response “Created topic topic_name” 5. Run the list command to verify if its created. 6. Now check the data directory of kakfa. There would not be any for the newly created topic. -- This message was sent by Atlassian JIRA (v6.3.4#6332)