使用Hive Sink时报错：“Unsupported Hive2 protocol”

[shens16]

问题简介：
我使用ClickHouse reader和Hive writer，可以成功提交flink job，但任务在flink中运行几秒钟后就显示任务failed。
查看flink错误日志发现如下错误：
`Caused by: org.apache.thrift.TException: Unsupported Hive2 protocol
    at org.apache.hive.jdbc.HiveConnection.openSession(HiveConnection.java:595)
    ... 32 more

    at com.dtstack.chunjun.connector.hive.util.HiveDbUtil.connect(HiveDbUtil.java:222)
    at com.dtstack.chunjun.connector.hive.util.HiveDbUtil.connect(HiveDbUtil.java:203)
    at com.dtstack.chunjun.connector.hive.util.HiveDbUtil$1.call(HiveDbUtil.java:96)
    at com.dtstack.chunjun.connector.hive.util.HiveDbUtil$1.call(HiveDbUtil.java:93)
    at com.dtstack.chunjun.util.RetryUtil$Retry.call(RetryUtil.java:141)
    at com.dtstack.chunjun.util.RetryUtil$Retry.doRetry(RetryUtil.java:80)
    ... 21 more
.
    at com.dtstack.chunjun.connector.hive.util.HiveDbUtil.getConnectionWithRetry(HiveDbUtil.java:104)
    at com.dtstack.chunjun.connector.hive.util.HiveDbUtil.getConnection(HiveDbUtil.java:86)
    at com.dtstack.chunjun.connector.hive.util.HiveUtil.createHiveTableWithTableInfo(HiveUtil.java:90)
    at com.dtstack.chunjun.connector.hive.sink.HiveOutputFormat.checkCreateTable(HiveOutputFormat.java:386)
    at com.dtstack.chunjun.connector.hive.sink.HiveOutputFormat.primaryCreateTable(HiveOutputFormat.java:357)
    at com.dtstack.chunjun.connector.hive.sink.HiveOutputFormat.openInternal(HiveOutputFormat.java:114)
    at com.dtstack.chunjun.sink.format.BaseRichOutputFormat.open(BaseRichOutputFormat.java:262)
    at com.dtstack.chunjun.connector.hive.sink.HiveOutputFormat.open(HiveOutputFormat.java:94)
    at com.dtstack.chunjun.sink.DtOutputFormatSinkFunction.open(DtOutputFormatSinkFunction.java:95)
    at org.apache.flink.api.common.functions.util.FunctionUtils.openFunction(FunctionUtils.java:34)
    at org.apache.flink.streaming.api.operators.AbstractUdfStreamOperator.open(AbstractUdfStreamOperator.java:102)
    at org.apache.flink.streaming.api.operators.StreamSink.open(StreamSink.java:46)
    at org.apache.flink.streaming.runtime.tasks.OperatorChain.initializeStateAndOpenOperators(OperatorChain.java:433)
    at org.apache.flink.streaming.runtime.tasks.StreamTask.lambda$beforeInvoke$2(StreamTask.java:545)
    at org.apache.flink.streaming.runtime.tasks.StreamTaskActionExecutor$SynchronizedStreamTaskActionExecutor.runThrowing(StreamTaskActionExecutor.java:93)
    at org.apache.flink.streaming.runtime.tasks.StreamTask.beforeInvoke(StreamTask.java:535)
    at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:575)
    at org.apache.flink.runtime.taskmanager.Task.doRun(Task.java:758)
    at org.apache.flink.runtime.taskmanager.Task.run(Task.java:573)
    at java.lang.Thread.run(Thread.java:745)`

环境信息：
操作系统：Ubuntu16.04.7
hive版本：2.3.9
flink版本：1.12.7
chunjun版本：1.12.7（最新）

我编写的测试用的json文件如下：
{ "job" : { "content" : [ { "reader" : { "name" : "clickhousereader", "parameter" : { "column":[ "packet_name","node_id","gateway_id" ], "connection" : [ { "jdbcUrl" : [ "jdbc:clickhouse://xxxx:8123/default" ], "schema": "bp001", "table" : [ "HOS_4000" ] }], "username" : "xxx", "password" : "xxxxxx" } }, "writer": { "name" : "hivewriter", "parameter" : { "jdbcUrl" : "jdbc:hive2://192.168.111.111:10000/chunjun", "username" : "", "password" : "", "fileType" : "text", "writeMode" : "overwrite", "charsetName" : "UTF-8", "tablesColumn" : "{\"chunjun.bp001\":[{\"type\":\"STRING\",\"key\":\"packet_name\"},{\"type\":\"STRING\",\"key\":\"node_id\"},{\"type\":\"STRING\",\"key\":\"gateway_id\"}]}", "partition" : "pt", "partitionType" : "MINUTE", "defaultFS" : "hdfs://ns", "hadoopConfig" : { "dfs.ha.namenodes.ns": "nn1,nn2", "fs.defaultFS": "hdfs://ns", "dfs.namenode.rpc-address.ns.nn2": "ip:9000", "dfs.client.failover.proxy.provider.ns": "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider", "dfs.namenode.rpc-address.ns.nn1": "ip:9000", "dfs.nameservices": "ns", "fs.hdfs.impl.disable.cache": "true", "hadoop.user.name": "root", "fs.hdfs.impl": "org.apache.hadoop.hdfs.DistributedFileSystem" } } } } ], "setting": { "speed": { "channel": 1 } } } }

补充说明：
1、flink-sql的客户端可成功连接操作hive，说明flink与hive的集成没有问题
2、hive的beeline客户端可正常使用，使用其它客户端通过json配置文件中的hive的jdbcURL也可以成功连接上hive