The only restrictions on indexing in MySQL Cluster are that text indexes and prefixes aren't supported. Did you try using USE INDEX or FORCE INDEX?
Why are you running four MySQL server nodes? Only one is necessary. In any case, you might want to verify the setup using just one SQL node before adding more.
In addition, the correct syntax is ENGINE=NDBCLUSTER -- TYPE is deprecated.
I'm not sure what you've got going on there, but it sounds like you've got a configuration issue whereby you've effectively four separate DBs, one running on each server -- IOW, that no clustering is actually taking place. Are you sure that the NDB Cluster engine is even available? Verify this using SHOW VARIABLES LIKE 'have_%'; and SHOW CREATE TABLE.
*ADDED* -- A couple more things I thought of:
1. Did you try running the count on a particular column rather than usng the wildcard?
2. If you're using multiple SQL nodes: Cluster databases aren't autodiscovered yet (we're working on this for a future release), so when you create a DB on one SQL node, you need to create it explicitly on all SQL nodes. Once you run the appropriate CREATE DATABASE statement, the existing tables in that DB should be automatically discovered.
Hope something I've said here proves helpful.
Jon Stephens
MySQL Documentation Team @ Oracle
MySQL Dev Zone
MySQL Server Documentation
Oracle
Edited 1 time(s). Last edit at 04/02/2005 12:05PM by Jon Stephens.