Hello, we are considering using gluster (and its replication feature) for an application that we are developing that will require high availability. Our requirement of gluster is that it would keep a replicated file system in synch when files are introduced from any particular node to its local system, and that it continue to provide read/write access when a companion server is down (say for maintenance). When the companion server comes back online, gluster must "catch up" with any changes to the file system. Is gluster appropriate for this use? How will a server be notified that a companion server is down? How will it be notified that a companion server is back online and has completed its synchronization? Thanks for any advice!
