Bug #69361 NDB tables get dropped and reinserted using MySQL Replication in MySQLCluster

• Submitted: 31 May 2013 6:40
• Modified: 19 May 2016 11:53
• Reporter: Shalva Cutler
• Status: Can't repeat
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S2 (Serious)
• Version: MySQL Cluster 7.2.12
• OS: Linux
• Assigned to: MySQL Verification Team
• CPU Architecture: Any

[31 May 2013 6:40] Shalva Cutler

Description:
When Using MySQL replication between two data nodes in MySQL Cluster the NDB tables are continuously added and dropped over and over. You can leave it for days and it continues to cycle this way. If you stop the slave it stops if you start the slave again it picks up again.

How to repeat:
Use these config files edit the directories and IP to your personal data:

Master:

http://pastebin.com/DxvM8VeG

Slave:

http://pastebin.com/dnaTKK0Y

Insert new NDB tables into the Master Database.

mysql -u root -p

both>use database;

both>show tables;

do the show tables command a few times and you will see the table count go up and down

slave>stop slave;

and you will see this stops

slave>start slave

It will begin again.

master>drop database;

master>create database;

Now in the master my.cnf comment out log-slave-updates

#    log-slave-updates

restart both instances.

slave>start slave;

Insert the new tables again and then check

both>use database;

both>show tables;

No more dropping and inserting tables

[19 May 2016 11:53] MySQL Verification Team

reproduced on 7.2.12, can't reproduce this on latest 7.2/7.4