On Thu, Oct 13, 2005 at 12:01:19PM -0700, Chris Travers wrote: > Stages 1 and 4 could be handled by Slony, while 2 and 3 would require > custom triggers. In essence this is really master/slave that appears > multimaster. You will have tradeoffs here in granularity of conflict > resolution versus performance. This sounds very similar to an idea Josh Berkus sketched for me while getting off a train at OSCON. So I bet I didn't get all the details :) But I think it's likely worth pursuing. > Why would Slony have to do your conflict resolution at all? Why not > just use it to replicate update logs and data sets, and leave conflict > resolution to custom triggers? Conflict resolution is really going to That would probably work. The real thing Slony can't stand is really long periods of disconnection. You might be able to hack the log shipping approach to help there, though. In any case, it's probably a subject for the Slony lists, once a full-ish proposal is ready. > Well, I want to thank Afilias for such a useful tool. Again, if we all Afilias didn't exactly contribute it -- it was always intended to be a community project, and we just contributed some seeds (read "Jan" and later "Chris") to get it growing. That seems to have worked -- I think there is now as much contribution from non-Afilias folks as Afilias folks, and I think Slony is better for it. A -- Andrew Sullivan | ajs@xxxxxxxxxxxxxxx Information security isn't a technological problem. It's an economics problem. --Bruce Schneier ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
