Q from Adrian: What is BI reporting? Is it reporting?
It stands for Business Inteligence which is a bit more than just reporting - also analysing the data
Q from Assaf: What's the max realistic number of nodes in my MySql cluster?
We mostly see up to 8 data nodes (remember that you add latency with more nodes). Regarding the SQL nodes we also see customers with more than 30 nodes
Q from Assaf: What is the difference between "data" and "SQL" nodes?
MySQL Server is the SQL node, the data nodes are the "HA-Storage Engine"
Q from Adrian: How can I config HA to execute a script during a failover? Is this is possible with V2? I want to check some constraints before the slave server takes over...
Yes, you can implement your own scripts which will be executed on failover
Q from Dave: Can you have multiple MySQL servers connected to a single HA storage engine - e.g. a SAN - if so, what best practices should be observed, e.g. read-only connections for all but one of the MySQL servers
A typical solution is a active/passive one - so you can only have one MySQL Server connected - We can use external locking for MyISAM to get around this but only for this Storage engine
Q: What about cluster on virtual nodes with LVS?
We do not recommend the use of MySQL Cluster on virtual servers. MySQL Cluster is a real-time system and it's optimised to get advantage from some specific hardware and firmware features. In a virtual environment you may have unexpected behaviours, hard to catch for our tech support team.
Q from Eric: All improvements for HA goes trough replication, which seems right. But what about performance: Is it only about to choose the right storage engine?
It is about choosing the right HA Solution for your requirements (asynchronous or synchronous replication, replication over a long distance, ...) and of course using the right storage engine and having the right configurations set