Re: Database replication method
| От | Yuksel Pinarbasi |
|---|---|
| Тема | Re: Database replication method |
| Дата | |
| Msg-id | 0891300009554A76B94ECFF01E11BCED@yukiPC обсуждение исходный текст |
| Ответ на | Re: Database replication method (Keith <keith@keithf4.com>) |
| Ответы |
Re: Database replication method
|
| Список | pgsql-novice |
Thanks Keith.
Those two branches will mostly be handling independent data.
Only time to time they will need to see a record that the other branch has been created.
I don't think there will be ant conflict. Because they will not read or edit the same row in a close time frame.
I don't think there will be ant conflict. Because they will not read or edit the same row in a close time frame.
It seems a Multi Master- Asynchronous replication will be the right solution in the end.
Anyway, I will start with a single database and work my way around later.
Yuksel
You're going down a rather complex path, especially for someone new to full RDBMS systems. Multi-master (as this is frequently called) is one of the most difficult problems to solve in database administration. A poor network between the two locations is going to make it even more difficult to manage. The conflict resolution for both sites editing the same data is more complex of a problem than I think you realize right now. If there's any chance you can stick with a single, more reliable offsite location that both locations can access, I'd highly recommend starting there, perhaps in Amazon EC2 or RDS. Then one site being down isn't a bottleneck to the other.Keith
В списке pgsql-novice по дате отправления: