Re: Configuration error: Error: Could not alloc node id at 172.xx.xx.2 port 1186: Connection done from wrong host ip 172.xx.x.4.

[Mauritz Sundell]

Hi Rait

It looks like you only have two API (mysqld) nodes defined in cluster and they are both in use (maybe by two mysqld).

One also need to have a few API nodes defined in cluster configuration to allow running most ndb tools, including ndb_restore.

How many extra API nodes depends of how many ndb tools you will run in parallel.

In your case I would add at least 5 API nodes in cluster configuration, which allows running a few ndb_restore plus some other ndb-tool in parallel.

This extra API nodes do not need to have a HostName, they can have if you want to restrict running tools from specific hosts.


After updating the config.ini, you need to restart ndb_mgmd with --reload option. And after that doing a rolling restart on whole cluster by restart all data nodes and mysql servers, see https://dev.mysql.com/doc/mysql-cluster-excerpt/8.0/en/mysql-cluster-rolling-restart.html.


After that SHOW should show something like:

[mysqld(API)] 7 node(s)
id=101 @172.xx.xx.5 (mysql-8.0.21 ndb-8.0.21)
id=102 (@172.xx.xx.6 (mysql-8.0.21 ndb-8.0.21)
id=103 (not connected, accepting connect from any host)
id=104 (not connected, accepting connect from any host)
...


Regards,
Mauritz