On 30 Jul 2009, at 6:19pm, Jay A. Kreibich wrote: > You're not really supposed to > dump and restore a multi-master system.
We're back to talking about synchronising different copies of the database again, aren't we ? Dumping and restoring a multi-master system means restoring the entire context of all the copies of the database. You have to not only restore the contents of one copy of the database, but also restore the state of all other copies. And if you're using a journaling method to manage your synchrony, restore the state of all the journals too. The only exception to this is if the dump was done when all sites agreed that they all had completely synchronised up-to-date copies of the database. And many multi-master databases are never in that state. So the question is 'what was dumped, and what state was it in' ? So the next question is how your journaling system was designed. You can design a journaling and synchronising system so that any site(s) can restore from a backup at any time, and the next time everyone synchronises they'll get their best data. But often this is not considered when the system is designed and the system loses data or worse still corrupts everyone else's. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
