On 18 Jul 2009, at 8:27pm, Ken wrote: > The "journal" file for sqlite is not a redo journal but rather an > undo journal. So it is not really possible to use the journal to > replicate.
I agree. > However, if one were to hack the code and open a redo file along > with the journal file. Then write the source blocks out as well. > This could then be copied and written against a target database on a > remote server. I know a bit about the problems of synching two copies of the same relational database. It is possible to do the job with a journal, but you need a journal at each site, they need more data in than SQLite journals, and you need some sort of intelligence about how to reconcile them. There's no "once size fits all" solution. To answer the original question: if only one site can make changes, the problem can be solved any number of ways, including rsync, ditto, making your own journal using TRIGGER, and many others. If more than one site can make changes then you're in for a world of hurt until you've read a few books or consulted someone with experience. Simon. _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
