HRegion 上的 doMiniBatchMutation 请求的行超出范围

Question

hbase客户端批量数据时出错。一开始还可以。过一段时间就错了！详细的错误是：

: 1 次，org.apache.hadoop.hbase.exceptions.FailedSanityCheckException: 请求的行超出了 HRegion idcard 上的 doMiniBatchMutation 的范围，bfef6945ac273d83\x00\x00\x00\x00\x00\x17\xCC$,1461584032622.dadb8843fe441dac4a3d4d76 ., startKey='bfef6945ac273d83\x00\x00\x00\x00\x00\x17\xCC$', getEndKey()='', row='9a6ec957205e1d74\x00\x00\x00\x00\x01\x90\x1F\xF5 ' 在 org.apache.hadoop.hbase.regionserver.RSRpcServices.doBatchOp(RSRpcServices.java:712) 在 org.apache.hadoop.hbase.regionserver.RSRpcServices.doNonAtomicRegionMutation(RSRpcServices.java:662) 在 org.apache.hadoop.hbase.regionserver.RSRpcServices.multi(RSRpcServices.java:2046) 在 org.apache.hadoop.hbase.protobuf.generated.ClientProtos$ClientService$2.callBlockingMethod(ClientProtos.java:32393) 在 org.apache.hadoop.hbase.ipc.RpcServer.call(RpcServer.java:2117) 在 org.apache.hadoop.hbase.ipc.CallRunner.run(CallRunner.java:104) 在 org.apache.hadoop.hbase.ipc.RpcExecutor.consumerLoop(RpcExecutor.java:133) 在 org.apache.hadoop.hbase.ipc.RpcExecutor$1.run(RpcExecutor.java:108) 在 java.lang.Thread.run(Thread.java:745)

环境是：

1. hbase hbase-1.1.3
2. hadoop2.6
3. hbase-client 1.2.0

hbase客户端的代码是：

public static void batchPutData(Connection connection, long startNum, long count) throws IOException, ParseException{
        //table
        Table table = connection.getTable(TableName.valueOf(TABLE_NAME));
        //index table
        Table index_table = connection.getTable(TableName.valueOf(INDEX_TABLE_NAME));

        //random name 
        RandomChineseName randomChineseName = new RandomChineseName();
        //random car
        RandomCar randomCar = new RandomCar();

        List<Put> puts = new ArrayList<Put>();
        List<Put> indexPlateputs = new ArrayList<Put>();
        for(long i = 0; i < count; i++){
            long index = startNum+i;

            Date birthdate = RandomUtils.randomDate();
            String birthdateStr =  DateUtil.dateToStr(birthdate, "yyyy-MM-dd");
            boolean isBoy = i%2==0?true:false;
            String name = isBoy?randomChineseName.randomBoyName():randomChineseName.randomGirlName();
            String nation = RandomUtils.randomNation();
            String plate = randomCar.randomPlate();

            byte[] idbuff = Bytes.toBytes(index); 
            String hashPrefix = MD5Hash.getMD5AsHex(idbuff).substring(0, 16);  

            //create a put for table
            Put put = new Put(Bytes.add(Bytes.toBytes(hashPrefix), idbuff));
            put.addColumn(Bytes.toBytes("idcard"), Bytes.toBytes("name"), Bytes.toBytes(name));
            put.addColumn(Bytes.toBytes("idcard"), Bytes.toBytes("sex"), Bytes.toBytes(isBoy?1:0));
            put.addColumn(Bytes.toBytes("idcard"), Bytes.toBytes("birthdate"), Bytes.toBytes(birthdateStr));
            put.addColumn(Bytes.toBytes("idcard"), Bytes.toBytes("nation"), Bytes.toBytes(nation));
            put.addColumn(Bytes.toBytes("idcard"), Bytes.toBytes("plate"), Bytes.toBytes(plate));
            puts.add(put);

            //create a put for index table
            String namehashPrefix = MD5Hash.getMD5AsHex(Bytes.toBytes(name)).substring(0, 16);  
            byte[] bprf = Bytes.add(Bytes.toBytes(namehashPrefix), Bytes.toBytes(name));
            bprf = Bytes.add(bprf, Bytes.toBytes(SPLIT), Bytes.toBytes(birthdateStr));
            Put namePut = new Put(Bytes.add(bprf, Bytes.toBytes(SPLIT), Bytes.toBytes(index)));
            namePut.addColumn(Bytes.toBytes("index"), Bytes.toBytes("idcard"), Bytes.toBytes(0));
            indexPlateputs.add(namePut);

            //insert for every ten thousands
            if(i%10000 == 0){
                table.put(puts);
                index_table.put(indexPlateputs);
                puts.clear();
                indexPlateputs.clear();
            }
        }
    }

Answer 1

似乎与 HBase 版本冲突。将 HBase 版本改为 1.1.4 或 1.0.0 或其他稳定版本试试。