Mysql gateway - connection getting closed

[rajitha703]

Hi,

I am facing two issues with the Mysql gateway :

The following query was fired twice from the mysql console :

mysql> select count(*) from supply where "2016-06-23 01:00:00" <= __time AND __time < "2016-06-23 02:00:00" group by countryId limit 10;
Query OK, 0 rows affected, 25971 warnings (0.10 sec)

mysql> select count() from supply where "2016-06-23 01:00:00" <= __time AND __time < "2016-06-23 02:00:00" group by countryId limit 10;
ERROR 2006 (HY000): MySQL server has gone away
No connection. Trying to reconnect...
Connection id: 21
Current database: ** NONE ***

Query OK, 0 rows affected, 25971 warnings (0.40 sec)

---

Issue 1. Notice that the first time , query runs fine but the second time, it says the Mysql server has gone away.
Following are the log lines on the Mysql server of Plyql. Both times I get the same error logged
Got SQL: select count(*) from supply where "2016-06-23 01:00:00" <= __time AND __time < "2016-06-23 02:00:00" group by countryId limit 10
Got SQL: SELECT DATABASE()
Error: Connection lost: The server closed the connection.

Issue 2 : The same query I tried to run directly using PLYQL as follows :
rajitha.r:~$ plyql -h mybroker:8082 -q 'select count(*) from supply where "2016-06-23 01:00:00" <= __time AND __time < "2016-06-23 02:00:00" group by countryId limit 10'
And this runs absolutely fine and gives the correct output. This is a bit strange as the Mysql gateway should also have returned the same output but doesn't.

I am using a latest version of Plyql. Can someone please help me on this.

[benvogan]

I haven't seen this with PlyQL but I have seen similar errors when there is a misconfiguration of the MySQL timeout and connection pooling on the client. If MySQL is timing out a connection that is maintained in a pool then when it is assigned you would see this error. Don't know if that is helpful in this case. Good luck!

[David Gee]

Please note that this project is now deprecated. Please see the Druid SQL docs for information on using SQL syntax with Druid.