impala-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From 俊杰陈 <cjjnj...@gmail.com>
Subject Re: Impala query seems doesn't return correct result when query after insert immediately.
Date Tue, 18 Apr 2017 07:06:38 GMT
Hi Alex
I found root cause. I thought I just started one backend according to
command I use, but actually I have another backend in other server which
always subscribe to my dev-server. As a result, my customized impala
backend fetch correct row, but backend impala did not retrieve rows.

Thanks for your reply!

2017-04-18 6:39 GMT+08:00 Alexander Behm <alex.behm@cloudera.com>:

> I tried to reproduce this issue, but without success.
>
> create database test;
> create table test.test2 (id int, name string, progress decimal(3,1))
> stored as parquet;
> insert into test.test2 values
> (3,'jj',12.1),
> (3,'test',15.3),
> (4,'jj',12.3),
> (2,'fs',15.2),
> (18,'jj',12.3),
> (13,'jj',12.3),
> (14,'jj',13.4),
> (2,'jj',12.1),
> (12,'jj',12.3),
> (6,'jj',12.3),
> (5,'jj',12.3),
> (20,'jj',12.3),
> (45,'jj',13.4),
> (15,'jj',15.5),
> (3,'jj',12.3),
> (1,'jj',23.5);
> create table test.test5 as select * from test.test2;
> select * from test.test5;
> select * from test.test5 where id=3;
>
> Since the issue is reproducible for you, can you post the complete series
> of steps? Thanks!
>
> Alex
>
> On Sun, Apr 16, 2017 at 6:03 PM, 俊杰陈 <cjjnjust@gmail.com> wrote:
>
>> Hi
>> Thanks for reply, I'm using incubator-impala from
>> https://github.com/apache/incubator-impala.git, and the commit
>> is a71636847fe742a9d0eb770516aff34ff16bbca1.
>>
>> The issue is reproducible.  Here is the show create table test5:
>> [bdpe822n2:21000] > show create table test5;
>> Query: show create table test5
>> +---------------------------------------------------------------------+
>> | result                                                              |
>> +---------------------------------------------------------------------+
>> | CREATE TABLE test.test5 (                                           |
>> |   id INT,                                                           |
>> |   name STRING,                                                      |
>> |   progress DECIMAL(3,1)                                             |
>> | )                                                                   |
>> | STORED AS PARQUET                                                   |
>> | LOCATION 'hdfs://bdpe822n2:20500/user/hive/warehouse/test.db/test5' |
>> +---------------------------------------------------------------------+
>>
>>
>> I don't use any load balancer, I started impala cluster with
>> bin/start-impala-cluster -s 1, and start impala shell with
>> bin/impala-shell.sh.
>>
>> 2017-04-14 23:37 GMT+08:00 Alexander Behm <alex.behm@cloudera.com>:
>>
>>> Thanks for posting. What Impala version are you running? Is the issue
>>> always reproducible? Are you behind a load balancer? Can you provide SHOW
>>> CREATE TABLE or the involved tables?
>>>
>>> Thanks!
>>>
>>> Alex
>>>
>>> On Fri, Apr 14, 2017 at 12:05 AM, 俊杰陈 <cjjnjust@gmail.com> wrote:
>>>
>>>> Hi
>>>> I met following strange issue, I insert a table from another table, and
>>>> execute immediately, it shows wrong result.
>>>> [bdpe822n2:21000] > insert into test.test5 select * from test.test2;
>>>>
>>>>  Query: insert into test.test5 select * from test.test2
>>>> Query submitted at: 2017-04-14 14:21:53 (Coordinator:
>>>> http://bdpe822n2:25000)
>>>> Query progress can be monitored at: http://bdpe822n2:25000/query_p
>>>> lan?query_id=ba42eeb19d262dc6:c10ef07500000000
>>>> Modified 16 row(s) in 5.22s
>>>> [bdpe822n2:21000] > select * from test5;
>>>> Query: select * from test5
>>>> Query submitted at: 2017-04-14 14:22:05 (Coordinator:
>>>> http://bdpe822n2:25000)
>>>> Query progress can be monitored at: http://bdpe822n2:25000/query_p
>>>> lan?query_id=7c4ff772ce33f171:6933975300000000
>>>> +----+------+----------+
>>>> | id | name | progress |
>>>> +----+------+----------+
>>>> | 3  | jj   | 12.1     |
>>>> | 3  | test | 15.3     |
>>>> | 4  | jj   | 12.3     |
>>>> | 2  | fs   | 15.2     |
>>>> | 18 | jj   | 12.3     |
>>>> | 13 | jj   | 12.3     |
>>>> | 14 | jj   | 13.4     |
>>>> | 2  | jj   | 12.1     |
>>>> | 12 | jj   | 12.3     |
>>>> | 6  | jj   | 12.3     |
>>>> | 5  | jj   | 12.3     |
>>>> | 20 | jj   | 12.3     |
>>>> | 45 | jj   | 13.4     |
>>>> | 15 | jj   | 15.5     |
>>>> | 3  | jj   | 12.3     |
>>>> | 1  | jj   | 23.5     |
>>>> +----+------+----------+
>>>> Fetched 16 row(s) in 0.18s
>>>> [bdpe822n2:21000] > select * from test5 where id=3;
>>>> Query: select * from test5 where id=3
>>>> Query submitted at: 2017-04-14 14:22:13 (Coordinator:
>>>> http://bdpe822n2:25000)
>>>> Query progress can be monitored at: http://bdpe822n2:25000/query_p
>>>> lan?query_id=23432b5caf4b1ab4:9ade99c300000000
>>>> +----+------+----------+
>>>> | id | name | progress |
>>>> +----+------+----------+
>>>> | 3  | jj   | 12.3     |
>>>> +----+------+----------+
>>>> Fetched 1 row(s) in 0.18s
>>>>
>>>> However, after I restart the impala cluster, it can get correct result.
>>>> Does anyone met issue like this? Should I need to run some synchronized
>>>> command?
>>>>
>>>>
>>>> --
>>>> Thanks & Best Regards
>>>>
>>>
>>>
>>
>>
>> --
>> Thanks & Best Regards
>>
>
>


-- 
Thanks & Best Regards

Mime
View raw message