[ https://issues.apache.org/jira/browse/HDFS-8479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14560103#comment-14560103 ]
Zhe Zhang commented on HDFS-8479: --------------------------------- Thanks Jing. I was about the include the following change: {code} /** * Decode based on the given input buffers and schema */ public static void decodeAndFillBuffer(final byte[][] decodeInputs, byte[] buf, AlignedStripe alignedStripe, int dataBlkNum, int parityBlkNum) { int[] decodeIndices = new int[parityBlkNum]; int pos = 0; for (int i = 0; i < alignedStripe.chunks.length; i++) { if (alignedStripe.chunks[i].state != StripingChunk.FETCHED && alignedStripe.chunks[i].state != StripingChunk.ALLZERO) { decodeIndices[pos++] = i; decodeInputs[i] = null; } } {code} Basically, HADOOP-11847 requires us to leave to-be-decoded slots as null. > Erasure coding: fix striping related logic in FSDirWriteFileOp to sync with > HDFS-8421 > ------------------------------------------------------------------------------------- > > Key: HDFS-8479 > URL: https://issues.apache.org/jira/browse/HDFS-8479 > Project: Hadoop HDFS > Issue Type: Sub-task > Reporter: Zhe Zhang > Assignee: Zhe Zhang > Fix For: HDFS-7285 > > Attachments: HDFS-8479-HDFS-7285.0.patch > > -- This message was sent by Atlassian JIRA (v6.3.4#6332)