Bug #28311 ndb_binlog_discover fails randomly

• Submitted: 8 May 2007 17:55
• Modified: 24 Aug 2010 7:34
• Reporter: Joerg Bruehe
• Status: Won't fix
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S3 (Non-critical)
• Version: mysql-5.1
• OS: Linux
• CPU Architecture: Any
• Tags: 5.1.18

[8 May 2007 17:55] Joerg Bruehe

Description:
Noted only once, during two test builds of 5.1.18 on all hosts:

-------------------------------------------------------
*** r/ndb_binlog_discover.result
--- r/ndb_binlog_discover.reject
***************
*** 12,17
--- 12,18
  master-bin.000002     #       Table_map       1       #       table_id: # (test.t1)
  master-bin.000002     #       Table_map       1       #       table_id: # (mysql.ndb_apply_status)
  master-bin.000002     #       Write_rows      1       #       table_id: #
+ master-bin.000002     #       Write_rows      1       #       table_id: #
  master-bin.000002     #       Write_rows      1       #       table_id: # flags: STMT_END_F
  master-bin.000002     #       Query   1       #       COMMIT
  drop table t1;
-------------------------------------------------------

I take it that "Write_rows" is the log of an Insert statement.
From the test case, the loop should terminate after the first non-failing insert -
so it seems like one insert got logged but still returned a non-zero code.

How to repeat:
Run it sufficiently often.