Bug #87610 Got error 741 'Unsupported alter table' from NDBCLUSTER
Submitted: 30 Aug 2017 14:49 Modified: 17 Oct 2017 16:35
Reporter: Jeff Steele Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S2 (Serious)
Version:7.5.4 OS:CentOS
Assigned to: CPU Architecture:Any
Tags: READ_BACKUP

[30 Aug 2017 14:49] Jeff Steele
Description:
I have a database with about 40 tables. I have been trying to alter the tables using the query: 

ALTER TABLE <table_name> algorithm=inplace, COMMENT="NDB_TABLE=READ_BACKUP=1"; 

This works on some tables, but in others I get the error: 

error 741 'Unsupported alter table' from NDBCLUSTER 

All the tables are ndbcluster tables and I can't find any rhyme or reason about why some tables are successfully altered and some aren't. This is with cluster version 7.5.4. 

How to repeat:
Run this query:

ALTER TABLE <table_name> algorithm=inplace, COMMENT="NDB_TABLE=READ_BACKUP=1"

On some tables it succeeds and on some it fails.
[30 Aug 2017 22:30] Jeff Steele
I upgraded to 7.5.7 and am getting the same results.
[30 Sep 2017 17:19] Mikael Ronström
This can happen after changing the number of log parts in a node.
The best thing to do here is to ensure that changing number of log parts requires an initial node
restart and that such a change will distribute the partitions as equally as possible among the
log parts.
[17 Oct 2017 16:35] MySQL Verification Team
Hi,

The bug is verified, you can use a workaround doing initial restart of all nodes (one by one so you don't lose your data) to solve the problem with problematic tables.

best regards
Bogdan