[ https://issues.apache.org/jira/browse/HDFS-8497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14570218#comment-14570218 ]
Li Bo commented on HDFS-8497: ----------------------------- Sorry for the late reply due to my annual leave from work. I run the latest code and still find the problem. An Exception is thrown: {{ org.apache.hadoop.HadoopIllegalArgumentException: Inputs not fully corresponding to erasedIndexes in null places at org.apache.hadoop.io.erasurecode.rawcoder.RSRawDecoder.doDecode(RSRawDecoder.java:132) at org.apache.hadoop.io.erasurecode.rawcoder.AbstractRawErasureDecoder.decode(AbstractRawErasureDecoder.java:77) at org.apache.hadoop.hdfs.server.datanode.erasurecode.ErasureCodingWorker$ReconstructAndTransferBlock.recoverTargets(ErasureCodingWorker.java:592) at org.apache.hadoop.hdfs.server.datanode.erasurecode.ErasureCodingWorker$ReconstructAndTransferBlock.run(ErasureCodingWorker.java:383) at java.lang.Thread.run(Thread.java:744) }} When we call {{decoder.decode}}, we should not allocate a buffer for the corrupt column, because decoder will check if the to-be-decoded column is null. > ErasureCodingWorker fails to do decode work > ------------------------------------------- > > Key: HDFS-8497 > URL: https://issues.apache.org/jira/browse/HDFS-8497 > Project: Hadoop HDFS > Issue Type: Sub-task > Reporter: Li Bo > Assignee: Li Bo > Attachments: HDFS-8497-HDFS-7285-01.patch > > > When I run the unit test in HDFS-8449, it fails due to the decode error in > ErasureCodingWorker. -- This message was sent by Atlassian JIRA (v6.3.4#6332)