drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Abhishek Girish (JIRA)" <j...@apache.org>
Subject [jira] [Created] (DRILL-2911) Queries fail with connection error when some Drillbit processes are down
Date Wed, 29 Apr 2015 21:28:06 GMT
Abhishek Girish created DRILL-2911:
--------------------------------------

             Summary: Queries fail with connection error when some Drillbit processes are
down
                 Key: DRILL-2911
                 URL: https://issues.apache.org/jira/browse/DRILL-2911
             Project: Apache Drill
          Issue Type: Bug
          Components: Execution - Flow
    Affects Versions: 0.9.0
            Reporter: Abhishek Girish
            Assignee: Chris Westin


Drill fails with connection error even when the Drill web UI also shows all drill-bits to
be up. However, some nodes do not list the Drillbit process. Looks like an inconsistent state.


Queries with simple scans execute successfully:
{code:sql}
select i_item_sk from item limit 5;
+------------+
| i_item_sk  |
+------------+
| 1          |
| 2          |
| 3          |
| 4          |
| 5          |
+------------+
5 rows selected (0.112 seconds)
{code}

Any query which might span across multiple drill-bits fails with connection error:
{code:sql}
SELECT 
* 
FROM     item i, 
                inventory inv
WHERE        inv.inv_item_sk = i.i_item_sk 
LIMIT 10;

Query failed: CONNECTION ERROR: Exceeded timeout while waiting send intermediate work fragments
to remote nodes.  Sent 4 and only heard response back from 3 nodes.
[5ada1a3e-d198-478b-941d-3c9bb917e494 on abhi7.qa.lab:31010]
Error: exception while executing query: Failure while executing query. (state=,code=0)
{code}

The issue could possibly be due to a previous failed query.  

Couldn't find the error code in logs. Have attached logs from all nodes for reference. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message