I get the subject error in the ndbd error log once I start the API process. An example is below. No error is reproted at the node until the API is started.

[root@db0 data]# tail db0.ispdone.internal.err 051222 9:02:21 [ERROR] /var/lib/mysql-cluster/bin/mysqld: Can't open file: 'radacct.ndb' (errno: 4009)

The ndb_mgm shows the ndb's seem to be stuck on 4 for a good long time, then stuck on 5 for a good long time, then zoom to 8 and disconnect. ndb_mgm> all status Node 2: starting (Phase 4) (Version 4.1.16) Node 3: starting (Phase 4) (Version 4.1.16)

Both nodes report the smae problem (error 4009).

A client connected to the API's shows this error

mysql> select * from radacct where 1; ERROR 1105 (HY000): Failed to open 'radacct', error while unpacking from engine

That may be because the API's never show getting connected to the cluster:

Cluster Configuration

--------------------- [ndbd(NDB)] 2 node(s) id=2 @10.10.0.50 (Version: 4.1.16, starting, Nodegroup: 0, Master) id=3 @10.10.0.51 (Version: 4.1.16, starting, Nodegroup: 0)

[ndb_mgmd(MGM)] 1 node(s) id=1 @10.10.0.10 (Version: 4.1.16)

[mysqld(API)] 2 node(s) id=4 (not connected, accepting connect from any host) id=5 (not connected, accepting connect from any host)

The radacct table is likely to be large but not to the extent to cause trouble I wouldn't think. Of course I can't determine how large without being able to query it. Any ideas? I urgently need some help on this as it is a live cluster.