Details
-
New Feature
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
2.2.0, 2.2.1
-
None
-
2017-49
Description
The following is an excerpt from a MaxScale log.
2017-12-19 19:00:10 info : (6) [readwritesplit] > Autocommit: [enabled], trx is [not open], cmd: (0x03) COM_QUERY, type: QUERY_TYPE_READ, stmt: select next value for seq1
|
2017-12-19 19:00:10 info : (6) [readwritesplit] Route query to slave [192.168.121.95]:3306 <
|
2017-12-19 19:00:53 info : (6) [readwritesplit] > Autocommit: [enabled], trx is [not open], cmd: (0x03) COM_QUERY, type: QUERY_TYPE_READ|QUERY_TYPE_WRITE, stmt: select nextval(seq1)
|
2017-12-19 19:00:53 info : (6) [readwritesplit] Route query to master [192.168.121.57]:3306 <
|
2017-12-19 19:03:30 info : (6) [readwritesplit] > Autocommit: [enabled], trx is [not open], cmd: (0x03) COM_QUERY, type: QUERY_TYPE_READ, stmt: select seq1.nextval
|
2017-12-19 19:03:30 info : (6) [readwritesplit] Route query to slave [192.168.121.95]:3306 <
|
It shows that the logically equivalent statements are not handled in a consistent way. All of the statements should be classified as reads that must be sent to the master, the same way that @@last_insert_id is treated.
Attachments
Issue Links
- relates to
-
MXS-2080 In SQL_MODE=ORACLE, sequence_name.nextval can get routed to slave
-
- Closed
-