[KYLIN-3930] ArrayIndexOutOfBoundsException when building

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: all
• Fix Version/s: v2.6.2
• Component/s: Job Engine
• Assignee: Unassigned
• Reporter: Jacky Woo
• Created: 02/Apr/19 02:17
• Updated: 18/May/19 02:00
• Resolved: 13/May/19 07:47

Description

ArrayIndexOutOfBoundsException when building.

I hive a cube building error with kylin-2.5.0:

2019-03-31 02:45:18,460 ERROR [main] org.apache.kylin.engine.mr.KylinMapper:
java.lang.ArrayIndexOutOfBoundsException
        at java.lang.System.arraycopy(Native Method)
        at org.apache.kylin.engine.mr.common.NDCuboidBuilder.buildKeyInternal(NDCuboidBuilder.java:106)
        at org.apache.kylin.engine.mr.common.NDCuboidBuilder.buildKey(NDCuboidBuilder.java:71)
        at org.apache.kylin.engine.mr.steps.NDCuboidMapper.doMap(NDCuboidMapper.java:112)
        at org.apache.kylin.engine.mr.steps.NDCuboidMapper.doMap(NDCuboidMapper.java:47)
        at org.apache.kylin.engine.mr.KylinMapper.map(KylinMapper.java:77)
        at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:146)
        at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:796)
        at org.apache.hadoop.mapred.MapTask.run(MapTask.java:342)
        at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:164)
        at java.security.AccessController.doPrivileged(Native Method)
        at javax.security.auth.Subject.doAs(Subject.java:422)
        at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1698)
        at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:158)

I checked the code of "NDCuboidBuilder.buildKeyInternal" method

private void buildKeyInternal(Cuboid parentCuboid, Cuboid childCuboid, ByteArray[] splitBuffers, ByteArray newKeyBodyBuf) {
        RowKeyEncoder rowkeyEncoder = rowKeyEncoderProvider.getRowkeyEncoder(childCuboid);

        // rowkey columns
        long mask = Long.highestOneBit(parentCuboid.getId());
        long parentCuboidId = parentCuboid.getId();
        long childCuboidId = childCuboid.getId();
        long parentCuboidIdActualLength = (long)Long.SIZE - Long.numberOfLeadingZeros(parentCuboid.getId());
        int index = rowKeySplitter.getBodySplitOffset(); // skip shard and cuboidId
        int offset = RowConstants.ROWKEY_SHARDID_LEN + RowConstants.ROWKEY_CUBOIDID_LEN; // skip shard and cuboidId
        for (int i = 0; i < parentCuboidIdActualLength; i++) {
            if ((mask & parentCuboidId) > 0) {// if the this bit position equals
                // 1
                if ((mask & childCuboidId) > 0) {// if the child cuboid has this
                    // column
                    System.arraycopy(splitBuffers[index].array(), splitBuffers[index].offset(), newKeyBodyBuf.array(), offset, splitBuffers[index].length());
                    offset += splitBuffers[index].length();
                }
                index++;
            }
            mask = mask >> 1;
        }

        rowkeyEncoder.fillHeader(newKeyBodyBuf.array());
    }

Found that "offset = SHARDID_LEN + CUBOIDID_LEN" , which is wrong when cube is not sharding. In my case my cube's storage type is 0, which means it is not sharding.

So, I set offset according to cube sharding, like below:

int offset = rowKeySplitter.getHeaderLength(); // skip shard and cuboidId

After modifying building succeeds in my environment.