Bug #14516 Restart of cluster can cause NDB API replication failure
Submitted: 31 Oct 2005 17:34 Modified: 26 Apr 2006 5:44
Reporter: Jonathan Miller Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:5.1.2 OS:Linux (Linux)
Assigned to: Tomas Ulin CPU Architecture:Any

[31 Oct 2005 17:34] Jonathan Miller 
Description:
During Cluster Replication stress testing, the NDB API started returning:
ERROR: 4006 Connect failure - out of connection objects (increase MaxNoOfConcurrentTransactions)
Status: Temporary error, Classification: Overload error
File: Bank.cpp (Line: 984)

The applications were stopped and the slave allowed to catch-up with master. Once the slave was caught up with master, the config.ini was updated and the cluster restarted without taking down any mysqld processes.

Once the master cluster was back up, all applications were restarted. Yet the BANK application did not continue to replicate.

In talking with development, NDB API have to have the "Show Table" done again once the master is restarted if you want replication to continue. 

How to repeat:
see above

Suggested fix:
Replication of these database once known to the master should be automatic and should not require users to run a "Show tables" command after restarting the master cluster.
[24 Apr 2006 19:36] 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/5425
[26 Apr 2006 5:44] Tomas Ulin 
no need to document, was fixed before by a previous patch