Description
2017-05-25 17:56:34,967 INFO [RpcServer.default.FPBQ.Fifo.handler=3,queue=0,port=50801] master.HMaster$11(2297): Client=hao//127.0.0.1 disable testModifyColumnFamily 2017-05-25 17:56:37,974 INFO [RpcClient-timer-pool1-t1] client.AsyncHBaseAdmin$TableProcedureBiConsumer(2219): Operation: DISABLE, Table Name: default:testModifyColumnFamily failed with Failed after attempts=3, exceptions: Thu May 25 17:56:35 CST 2017, , java.io.IOException: Call to localhost/127.0.0.1:50801 failed on local exception: org.apache.hadoop.hbase.ipc.CallTimeoutException: Call id=294, waitTime=1008, rpcTimeout=1000 Thu May 25 17:56:37 CST 2017, , java.io.IOException: Call to localhost/127.0.0.1:50801 failed on local exception: org.apache.hadoop.hbase.ipc.CallTimeoutException: Call id=295, waitTime=1299, rpcTimeout=1000 Thu May 25 17:56:37 CST 2017, , java.io.IOException: Call to localhost/127.0.0.1:50801 failed on local exception: org.apache.hadoop.hbase.ipc.CallTimeoutException: Call id=296, waitTime=668, rpcTimeout=660 017-05-25 17:56:38,936 DEBUG [RpcServer.default.FPBQ.Fifo.handler=3,queue=0,port=50801] procedure2.ProcedureExecutor(788): Stored procId=15, owner=hao, state=RUNNABLE:DISABLE_TABLE_PREPARE, DisableTableProcedure table=testModifyColumnFamily
For this disable table procedure, master return the procedure id when it submit the procedure to ProcedureExecutor. And the above procedure take 4 seconds to submit. So the disable table call failed because the rpc timeout is 1 seconds and the retry number is 3.
For admin operation, I thought we don't need change the default timeout config in unit test. And the retry is not need, too. (Or we can set a retry > 1 to test nonce thing). Meanwhile, the default timeout is 60 seconds. So the test type may need change to LargeTests.