MySQL Bugs: #12880: MaxNoOfTables option applies incorrect

Bug #12880	MaxNoOfTables option applies incorrect
Submitted:	30 Aug 2005 14:49	Modified:	11 Apr 2007 11:22
Reporter:	Serge Kozlov	Email Updates:
Status:	Verified	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S3 (Non-critical)
Version:	5.0+	OS:	Linux (Mandrake Linux 10)
Assigned to:		CPU Architecture:	Any
Tags:	5.0.11-max

Description:
I add 'MaxNoOfTables=10' option to '[ndb default]' section and run a testcase which try to create 100 tables. They are created successfully though amount of tables is limited in config.

How to repeat:
1. Add 'MaxNoOfTables=10' option to '[ndb default]' section in ndb/ndb_config_2_node.ini

2. Create short testcase like following:

--source include/have_ndb.inc
--source include/not_embedded.inc

--disable_warnings
drop table if exists t1;
--enable_warnings

let $1 = 100;
while ($1)
{
    eval create table test.t$1 (a char(1)) engine=ndb;
    dec $1;
}

let $1 = 100;
while ($1)
{
    eval drop table if exists test.t$1;
    dec $1;
}

3. Run it

Documentation update needed:

the parameter MaxNoTables is not strictly enforced, it is meerly used as a hint from the user to know how many tables he will be using, to setup some preallocated stuctures.

Same is true for other paramenters as MaxNoOrderedIndexes and MaxNoAttributes

the parameters are optional to specify (they have default values)

_and_ they are not strictly enforced, they are meant to tell the "cluster" that the application is planning to use this number as a max.  And correct operation is not guarenteed if they are exceeded, however no checks are done that they aren't.

Thank you for your bug report. This issue has been addressed in the
documentation. The updated documentation will appear on our website
shortly, and will be included in the next release of the relevant
product(s).

Additional info:

Cluster chapter of Manual updated with info from Tomas.

reopened, this should at least print a warning when exceeding the configured value

This should have been changed to a Docs bug when it was filed. Doing so now,
reassigning to myself.

I'm giving this one back to the Cluster team, because 

(a) We have behaviour that is seemingly at odds with what is intended

(b) There's a request to supply warnings or errors