I am trying to setup a multi-master replication architecture and wanted to post my thoughts to see if I am going about it the right way and to see if anyone can suggest problems with this architecture before I start.
I have 3 "local" sites and one central site (AWS instance). If a user is at a local site, they need to make updates to the local database. If they are traveling, they will not access the local site, but rather the central site. The local sites may not have access to the internet for days or weeks at a time (one site is in a very remote location). Note that there is NO failover from one site to another, this is just for data replication purposes to make sure everyone is working with the same data.
Will mySQL replication handle the above multi-master scenario? My thinking is that this would be a star type of topology with the AWS instance being the central node and the 3 remote sites replicating to the center and the center then replicating those changes out to the other remote sites.
What about conflict resolution? I doubt there will be many conflicts since each site is really responsible for updates of their own data, but there is still the possibility of some conflicts, especially with data written to the central node. However, how are conflicts handled with mySQL replication?
I have 3 "local" sites and one central site (AWS instance). If a user is at a local site, they need to make updates to the local database. If they are traveling, they will not access the local site, but rather the central site. The local sites may not have access to the internet for days or weeks at a time (one site is in a very remote location). Note that there is NO failover from one site to another, this is just for data replication purposes to make sure everyone is working with the same data.
Will mySQL replication handle the above multi-master scenario? My thinking is that this would be a star type of topology with the AWS instance being the central node and the 3 remote sites replicating to the center and the center then replicating those changes out to the other remote sites.
What about conflict resolution? I doubt there will be many conflicts since each site is really responsible for updates of their own data, but there is still the possibility of some conflicts, especially with data written to the central node. However, how are conflicts handled with mySQL replication?