56b411aea9
All workers share an epoll instance that is added level-triggered to the epoll instance of each Worker. This is intended to be used together with listening sockets. When a listening socket is added to the shared epoll instance the effect is that EPOLLIN will be active for it whenever there is a connection pending on a listening socket added to that epoll instance. When that occurs all workers in their epoll_wait()-calls will return. When the workers subsequently call epoll_wait() on the shared epoll instance, that will return with an event provided some other thread(s) has not yet called accept() on the listening socket. As each worker extracts just one event at a time and calls accept just once before calling epoll_wait(), it means that the client connections will be distributed evenly across all workers, provided the load on the workers is roughly the same. If it isn't then a worker with less load will get more connections to handle (which will even out the load).
MaxScale by MariaDB Corporation
The MariaDB Corporation MaxScale is an intelligent proxy that allows forwarding of database statements to one or more database servers using complex rules, a semantic understanding of the database statements and the roles of the various servers within the backend cluster of databases.
MaxScale is designed to provide load balancing and high availability functionality transparently to the applications. In addition it provides a highly scalable and flexible architecture, with plugin components to support different protocols and routing decisions.
MaxScale is implemented in C and makes extensive use of the asynchronous I/O capabilities of the Linux operating system. The epoll system is used to provide the event driven framework for the input and output via sockets.
The protocols are implemented as external shared object modules which can be loaded at runtime. These modules support a fixed interface, communicating the entries points via a structure consisting of a set of function pointers. This structure is called the "module object".
The code that routes the queries to the database servers is also loaded as external shared objects and are referred to as routing modules.
An Google Group exists for MaxScale that can be used to discuss ideas, issues and communicate with the MaxScale community.
We're also on the #maria and #maxscale channels on FreeNode.
Please report all feature requests, improvements and bugs in the MariaDB Jira.
Documentation
For information about installing and using MaxScale, please refer to the documentation. The official documentation can be found on the MariaDB Knowledge Base.
The module and configuration documentation can be found in the Documentation directory of the source tree.
Contributing Code
Read the Contributing page on the wiki for more information on how to do pull request and where to do them.