Fri Mar 9 08:06:20 PST 2007
- Previous message: [Slony1-general] Architectural guidance - cascaded replication
- Next message: [Slony1-general] slon documentation bugs and annoyances
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Marco Gaiarin wrote: > As you ave seen i'm still a novice on slony, but i've just some doubt > on it. > > 1) master have all the connect information for the slave(s), so why > there's a slon daemon also on the slave? master can simply > insert/delete/update data on their own? > > As far as the slon daemon is concerned, there is no such thing as a "master" or a "slave." There is just: 1. The node that I am managing, and 2. Other nodes that I draw events from. The result is that switching roles (e.g. - when you do MOVE SET or FAILOVER) does not require shifting processes around. When a node takes over the "provider" role for a set, there's nothing too special about that - the slon managing the node can already manage that. > 2) (probably conflicting with 1) why on slave there's no a trigger that > simply redirect insert/delete/update on the master, instead of > preventing it? > That would mean building a multimaster replication system, and conflict resolution is a Big Problem with doing that. Slony-I was designed as a single-master replication system; let it be good at what it is, rather than trying to force Big Problems onto it...
- Previous message: [Slony1-general] Architectural guidance - cascaded replication
- Next message: [Slony1-general] slon documentation bugs and annoyances
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list