[ 
https://issues.apache.org/jira/browse/DRILL-6251?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Vitalii Diravka updated DRILL-6251:
-----------------------------------
    Description: 
On CentoOS cluster Drill hangs while querying sys tables after "use dfs;" 
(embedded or distributed mode):
{code}
0: jdbc:drill:> select * from sys.version;
+----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
| version  |                 commit_id                 |                    
commit_message                    |        commit_time         |    build_email 
    |         build_time         |
+----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
| 1.13.0   | 796fcf051b3553c4597abbdca5ca247b139734ba  | [maven-release-plugin] 
prepare release drill-1.13.0  | 13.03.2018 @ 11:39:14 IST  | par...@apache.org  
| 13.03.2018 @ 13:13:45 IST  |
+----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
1 row selected (3.784 seconds)
0: jdbc:drill:> use dfs;
+-------+----------------------------------+
|  ok   |             summary              |
+-------+----------------------------------+
| true  | Default schema changed to [dfs]  |
+-------+----------------------------------+
1 row selected (0.328 seconds)
0: jdbc:drill:> select * from sys.version;
Error: Statement canceled (state=,code=0)
0: jdbc:drill:> 
{code}
*Note*: there is no failure on local debian machine with Drill in embedded mode.
dfs pugin configs are default (with "connection": "file:///", other file 
systems works good).

This failure is connected to DRILL-5089 and Calcite rebase. 
Related commits:
https://github.com/apache/drill/commit/450e67094eb6e9a6484d7f86c49b51c77a08d7b2
https://github.com/apache/drill/commit/18a71a38f6bd1fd33d21d1c68fc23c5901b0080a

After analyzing in remote debug I found the following flow:
"dfs" DynamicRootSchema is created, then a new "sys" one is created.
After Calcite validate "sys" SimpleCalciteSchema is created. But in 
WorkspaceSchemaFactory#create  wrong WorkspaceConfig is left and "/" is 
combined with "sys".

  was:
On CentoOS cluster Drill hangs while querying sys tables after "use dfs;" 
(embedded or distributed mode):
{code}
0: jdbc:drill:> select * from sys.version;
+----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
| version  |                 commit_id                 |                    
commit_message                    |        commit_time         |    build_email 
    |         build_time         |
+----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
| 1.13.0   | 796fcf051b3553c4597abbdca5ca247b139734ba  | [maven-release-plugin] 
prepare release drill-1.13.0  | 13.03.2018 @ 11:39:14 IST  | par...@apache.org  
| 13.03.2018 @ 13:13:45 IST  |
+----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
1 row selected (3.784 seconds)
0: jdbc:drill:> use dfs;
+-------+----------------------------------+
|  ok   |             summary              |
+-------+----------------------------------+
| true  | Default schema changed to [dfs]  |
+-------+----------------------------------+
1 row selected (0.328 seconds)
0: jdbc:drill:> select * from sys.version;
Error: Statement canceled (state=,code=0)
0: jdbc:drill:> 
{code}
*Note*: there is not failure on local debian machine in embedded mode.
dfs pugin configs are default (with "connection": "file:///", other file 
systems works good).

This failure is connected to DRILL-5089 and calcite rebase. 
Related commits:
https://github.com/apache/drill/commit/450e67094eb6e9a6484d7f86c49b51c77a08d7b2
https://github.com/apache/drill/commit/18a71a38f6bd1fd33d21d1c68fc23c5901b0080a

After analyzing in remote debug I found the following flow:
"dfs" DynamicRootSchema is created, then a new "sys" one is created.
After calcite validate "sys" SimpleCalciteSchema is created. But in 
WorkspaceSchemaFactory#create  wrong WorkspaceConfig is left and "/" is 
combined with "sys".


> Queries from system tables are hang
> -----------------------------------
>
>                 Key: DRILL-6251
>                 URL: https://issues.apache.org/jira/browse/DRILL-6251
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Execution - Flow
>    Affects Versions: 1.13.0
>            Reporter: Vitalii Diravka
>            Priority: Major
>
> On CentoOS cluster Drill hangs while querying sys tables after "use dfs;" 
> (embedded or distributed mode):
> {code}
> 0: jdbc:drill:> select * from sys.version;
> +----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
> | version  |                 commit_id                 |                    
> commit_message                    |        commit_time         |    
> build_email     |         build_time         |
> +----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
> | 1.13.0   | 796fcf051b3553c4597abbdca5ca247b139734ba  | 
> [maven-release-plugin] prepare release drill-1.13.0  | 13.03.2018 @ 11:39:14 
> IST  | par...@apache.org  | 13.03.2018 @ 13:13:45 IST  |
> +----------+-------------------------------------------+------------------------------------------------------+----------------------------+--------------------+----------------------------+
> 1 row selected (3.784 seconds)
> 0: jdbc:drill:> use dfs;
> +-------+----------------------------------+
> |  ok   |             summary              |
> +-------+----------------------------------+
> | true  | Default schema changed to [dfs]  |
> +-------+----------------------------------+
> 1 row selected (0.328 seconds)
> 0: jdbc:drill:> select * from sys.version;
> Error: Statement canceled (state=,code=0)
> 0: jdbc:drill:> 
> {code}
> *Note*: there is no failure on local debian machine with Drill in embedded 
> mode.
> dfs pugin configs are default (with "connection": "file:///", other file 
> systems works good).
> This failure is connected to DRILL-5089 and Calcite rebase. 
> Related commits:
> https://github.com/apache/drill/commit/450e67094eb6e9a6484d7f86c49b51c77a08d7b2
> https://github.com/apache/drill/commit/18a71a38f6bd1fd33d21d1c68fc23c5901b0080a
> After analyzing in remote debug I found the following flow:
> "dfs" DynamicRootSchema is created, then a new "sys" one is created.
> After Calcite validate "sys" SimpleCalciteSchema is created. But in 
> WorkspaceSchemaFactory#create  wrong WorkspaceConfig is left and "/" is 
> combined with "sys".



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to