This shows you the differences between two versions of the page.
— | wiki:old:gsoc_proposal_fault_tolerance [2013/07/10 22:55] (current) – created - external edit 127.0.0.1 | ||
---|---|---|---|
Line 1: | Line 1: | ||
+ | ===== Fault Tolerance ===== | ||
+ | |||
+ | We would like to make OAR fault tolerant. For implanting such a solution, two nerve centers have to be taken in account and be reinforced: | ||
+ | |||
+ | * the database | ||
+ | * the OAR server | ||
+ | |||
+ | One part of the solution could be to set a monitoring system between the OAR server and an other spare server (by heartbeat system or any other watch method, see "this page": | ||
+ | |||
+ | For the database fault tolerance, replication mono master / mono slave seems to be a good solution in our context (see "MySQL replication": | ||
+ | |||
+ | The system has to work with both Postgres and Mysql databases. However this solution is not necessary the best regarding the overload if the database is much solicited. | ||
+ | |||
+ | The work of the intern will be to set benchmarks to test if database (Postgres and Mysql) replication is acceptable and to suggest other solutions. | ||
+ | |||
+ | At the end of this study, the intern will have to choose and implement a solution for both server and database fault tolerance. | ||
+ | |||
+ | The intern will be in constant relation with the team and will regularly discuss the work progression by audioconference and by mail. | ||
+ | |||
+ | ===== Classification ===== | ||
+ | |||
+ | Medium | ||
+ | |||
+ | ===== Intern required skills ===== | ||
+ | |||
+ | * Perl | ||
+ | * Fault tolerance on Linux servers | ||
+ | * Databases: Postgres, Mysql | ||
+ | * Linux | ||
+ | |||
+ | ===== Mentors ===== | ||
+ | * Joseph Emeras | ||
+ | * Bruno Bzeznik | ||