I'm trying to understand whether setting up a MySQL cluster running on Windows 2000/2003 is possible.
I am not a cluster expert, actually I have only read the MS white papers about it and done some Google search, but still not tried to do it in practice, so forgive me if I make silly questions. So far, I have gathered the following information. Windows 2000/2003 supports clustering in two forms: Cluster service (for high availability, i.e. failover) and Load Balancing. Currently I am only interested in failover. I have read that a typical configuration would be two identical servers, with a shared RAID disk array. In an active-passive configuration, only one of the servers would accept client requests and access the shared data. The other would be idle. If the first server failed, the second one would become active. I have also read that in order to fully support failover, applications would need special resource DLLs, written according to MS standards, which would be used to exchange messages with the cluster service. However, the documentation mentioned a generic resource DLL, which could be used to manage cluster unaware applications in a cluster configuration. Has anyone here ever tried to use this generic resource DLL? Would it work also for services (e.g. MySQL)? In theory, the service on the secondary cluster node would have to be started/stopped according to failover, else it could lead to access conflicts to the shared data. I also have a server application which would need to be on the cluster. It does not need independent data storage (it uses MySQL). Would this application be correctly started/stopped by the generic resource DLL? On MySQL website, I saw that a new product, MySQL Cluster, is in the works, but there is not much information about it (apart from the fact that it will be covered in the MySQL Conference). Do anybody know whether it will be available for Windows? Has an approximate release date been defined? I have already read about Replication, but as far as I know it has some problems. First, it is not transparent to the clients: even if you have a master and a slave which could take the master's place in case of failure, its IP would not be the same. Second, I would have to write the script to switch the master in case of failure. Third, this would solve the high availability problem for the database, but not for my server application... Thanks for any help. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]