Thanks for all the help so far, and sorry that I am still not sure that I am quite there yet.
I wanted to test if my cluster is working, and when I run a "show" on my mngt cluster, it shows that all nodes (2xdata + 1xsql) are connected.
I then stopped all nodes, and started ONLY the SQL node. It started fine.
I do understand that the data nodes need to be started first, but I just wanted to test something.
After I got the SQL node running, I then created a new schema, and thereafter a table.
However, I would think that this shouldn't be possible, since it wouldn't be able to firstly link to the mngt node (since it wasn't running), and then where exactly was it creating the schema, since the data nodes weren't running either!
When I checked the type of the table created, it was INNODB. Which again raises questions, since it should be NDB (or is this table type not an automatic selection to NDB?)
Perhaps someone can comment on this test I did?
How can I be sure that assuming all nodes have started, that when I create a new table that it gets stored on the DATA nodes, and not on the SQL node (which I'm guessing is what is happening, since the real data nodes weren't even running!
I wanted to test if my cluster is working, and when I run a "show" on my mngt cluster, it shows that all nodes (2xdata + 1xsql) are connected.
I then stopped all nodes, and started ONLY the SQL node. It started fine.
I do understand that the data nodes need to be started first, but I just wanted to test something.
After I got the SQL node running, I then created a new schema, and thereafter a table.
However, I would think that this shouldn't be possible, since it wouldn't be able to firstly link to the mngt node (since it wasn't running), and then where exactly was it creating the schema, since the data nodes weren't running either!
When I checked the type of the table created, it was INNODB. Which again raises questions, since it should be NDB (or is this table type not an automatic selection to NDB?)
Perhaps someone can comment on this test I did?
How can I be sure that assuming all nodes have started, that when I create a new table that it gets stored on the DATA nodes, and not on the SQL node (which I'm guessing is what is happening, since the real data nodes weren't even running!