spark-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andy Zhao <>
Subject Spark streaming lost data when ReceiverTracker writes Blockinfo to hdfs timeout
Date Tue, 26 Jul 2016 09:45:50 GMT
Hi guys, 

    I wrote a spark streaming program which consume 1000 messages from one
topic of Kafka, did some transformation, and wrote the result back to
another topic. But only found 988 messages in the second topic. I checked
log info and confirmed all messages was received by receivers. But I found a
hdfs writing time out message printed from Class BatchedWriteAheadLog. 
    I checkout source code and found code like this: 
    /** Add received block. This event will get written to the write ahead
log (if enabled). */ 
  def addBlock(receivedBlockInfo: ReceivedBlockInfo): Boolean = { 
    try { 
      val writeResult = writeToLog(BlockAdditionEvent(receivedBlockInfo)) 
      if (writeResult) { 
        synchronized { 
          getReceivedBlockQueue(receivedBlockInfo.streamId) +=
        logDebug(s"Stream ${receivedBlockInfo.streamId} received " + 
          s"block ${receivedBlockInfo.blockStoreResult.blockId}") 
      } else { 
        logDebug(s"Failed to acknowledge stream
${receivedBlockInfo.streamId} receiving " + 
          s"block ${receivedBlockInfo.blockStoreResult.blockId} in the Write
Ahead Log.") 
    } catch { 
      case NonFatal(e) => 
        logError(s"Error adding block $receivedBlockInfo", e) 

    It seems that ReceiverTracker tries to write block info to hdfs, but the
write operation time out, this cause writeToLog function return false, and 
this code "getReceivedBlockQueue(receivedBlockInfo.streamId) +=
receivedBlockInfo" is skipped. so the block info is lost. 

   The spark version I use is 1.6.1 and I did not turn on
   I want to know whether or not this is a designed behaviour. 


View this message in context:
Sent from the Apache Spark User List mailing list archive at

To unsubscribe e-mail:

View raw message