Fwd: Db synch - need advice - Mailing list pgsql-novice

Craig,

Thanks a bunch for the pointer I'll jump onto it after a few hours of sleep (it's just over midnight here). 

Thanks again for your help. If you are interested I can report back in the near future if this project goes anywhere.

Didier

On 8 January 2011 23:10, Craig Barnes <cjbarnes18@gmail.com> wrote:

Hi Didier,

Fair enough, this was the better use case for the dbase based system that I used back then.  That was implemented within the application using a change log and complex rules for duplicate checking.

I just recalled a replication methods reference table in the postgresql manual.  According to this, Multi Master replication is possible using Bacardo.  Looks like an interesting project, worth looking at it to see if it could fulfil your needs.

Regards

Craig

On 8 January 2011 19:44, Didier Gasser-Morlay <didiergm@gmail.com> wrote:

Craig,

Thanks a lot for the insight, I am fully aware of the risks. that being said, in THIS instance, I believe it can work because of the low volume of updates. It is indeed a new system in which I can do pretty much what I decide/need to do. This is why I am inquiring for strategy suggestions or possibly existing tools (one can always dream:))

As said previously, online access would be a preferred option (even a simple vpn would do!) but decent acces speed is not guaranteed by a long stretch! the sales guys are operating in fairly remote area (even lower mountains) where the GSM 3G is not available and won't be for a few years at best. hence the need to collect as much data as possible and sync when a connection is available.

Thanks again

Didier

On 8 January 2011 19:48, Craig Barnes <cjbarnes18@gmail.com> wrote:

Hi Didier,

A few years ago I worked with a Commercial Off The Shelf CRM application that supported  Database Synchronisation.  It worked but was prone to all sorts of difficulties including data loss and complicated conflict resolution processes.  Relational databases just don't lend themselves to this kind of work-flow.

My personal opinion would be to advise against synchronisation in this way in favour of online access.

Non relational databases work better with offline data, and there is some implementation in in Evolution, but I don’t think this allows for separate data sets.

Is there an existing system in place that wants to be augmented, or are you looking for a new system to implement?

Hope This Helps.

Craig