Mon May 12 10:18:49 PDT 2008
- Previous message: [Slony1-general] Origin/Provider understanding issue
- Next message: [Slony1-general] Subscriber nodes, full mesh or not
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Michael Gruetzner <mgruetzn at HTWM.De> writes: > I've read the documentation item 5.5 at > http://slony.info/documentation/concepts.html . > If I understand this correctly, it would not work, if a user updates a > table belonging to > a specific replication set on a node which is _not_ the origin of that > replication set. Correct; you cannot do updates on a non-origin node because Slony-I adds in a trigger using a function called deny_access() which prevents processes other than slon from doing Insert/Update/Delete. > Would it be possible to define multiple origin nodes for one > replication set to allow applications to update the same table on > multiple nodes? You're asking for multimaster replication; Slony-I was *not* designed as a multimaster replication system. There is some thinking taking place as to how to describe some limited sorts of update scenarios where it might be made permissible to do something like that, but that's at the "we don't really have useful thoughts yet" stage. -- select 'cbbrowne' || '@' || 'linuxfinances.info'; http://linuxfinances.info/info/ Hard work pays off in the long run. Laziness pays off now.
- Previous message: [Slony1-general] Origin/Provider understanding issue
- Next message: [Slony1-general] Subscriber nodes, full mesh or not
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list