Bug #17728 tests that fails are: ndb_gis rpl_ndb_multi_update2
Submitted: 26 Feb 2006 20:18 Modified: 14 Mar 2006 7:38
Reporter: Michael Widenius Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Replication Severity:S1 (Critical)
Version:5.1.8-BK OS:Linux (Linux)
Assigned to: Magnus Blåudd CPU Architecture:Any

[26 Feb 2006 20:18] Michael Widenius
Description:
make test fails in 5.1 when compiled with BUILD/compile-pentium-debug-max

The tests that fails are: ndb_gis rpl_ndb_multi_update2

ndb_gis fails by not producing warnings of type "Error 1538    Table storage engine 'ndbcluster' does not support the create option 'Binlog of table with BLOB attribute and no PK'

rpl_ndb_multi_update2 fails with:

mysqltest: In included file "./extra/rpl_tests/rpl_multi_update2.test": At line
35: could not sync with master ('select master_pos_wait('master-bin.000001', 131
1)' returned NULL)

The problem seams to be related to statement based replication

Another problem is that ndb_gis (and some other ndb tests) generates a lot of "memory not free'd" warnings.

How to repeat:
BUILD/compile-pentium-debug-max
make test

or

mysql-test-run --force --mysqld=--binlog-format=statement  ndb_gis rpl_ndb_multi_update2

For problem with lost memory, do:

 mysql-test-run --force t/ndb_gis.test 

And examine var/log/master.err for warnings
[27 Feb 2006 11:34] Valeriy Kravchuk
Thank you for a problem report. Verified just as described on Linux with 5.1.8-BK (ChangeSet@1.2184, 2006-02-27 11:21:20+04:00)
[1 Mar 2006 9:38] Magnus Blåudd
This happens all the time on my FC4 machine. Can't see it in pushbuild.
[1 Mar 2006 13:33] Magnus Blåudd
The ndb_gis test fails because of missing warnings. Those warnings are only displayed if the  flag "ndb_binlog_running" is set. Since the ndb binlog only support row based logging, the flag will only be set if --binlog_format=row. 

Since the testcase is not supposed to test binlogging but rather the gis datatypes I would suggest adding a pk to the tables used.
[1 Mar 2006 14:41] Bugs System
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/3330
[1 Mar 2006 14:54] Magnus Blåudd
Continue to work with the "rpl_ndb_multi_update" problem.
[1 Mar 2006 15:25] Magnus Blåudd
It seems like the slave thread terminates itself with "Unknown error 1105". See below excerpt from slave.log

060301 18:10:13 [Note] next log '/home/msvensson/mysql/bug17728/my51-bug17728/mysql-test/var/log/slave-relay-bin.000003' is currently active
060301 18:10:14 [Note] NDB Binlog: CREATE TABLE Event: REPL$test/t1
060301 18:10:15 [Note] NDB Binlog: CREATE TABLE Event: REPL$test/t2
060301 18:10:15 [ERROR] Slave: Error 'Can't write; duplicate key in table 't2'' on query. Default database: 'test'. Query: 'INSERT INTO t2 VALUES (NULL, 0), (NULL,1)', Error_code: 1022
060301 18:10:15 [Warning] Slave: Can't write; duplicate key in table 't2' Error_code: 1022
060301 18:10:15 [Warning] Slave: Unknown error Error_code: 1105
060301 18:10:15 [ERROR] Error running query, slave SQL thread aborted. Fix the problem, and restart the slave SQL thread with "SLAVE START". We stopped at log 'master-bin.000001' position 927
[3 Mar 2006 10:53] Bugs System
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/3421
[3 Mar 2006 13:47] Magnus Blåudd
Still no patch for rpl_ndb_multi_update2
[14 Mar 2006 7:37] Magnus Blåudd
rpl_ndb_multi_update problem will be fixed by bug#18208