> William Yu wrote:
>> Johnny Ljunggren wrote:
>>
>>> 1. Replication - multimaster
>>> I'll try to explain the setup to the best of my ability:
>>> Three centers:
>>> Main center - database with a backup database
>>> Center 1 - database with a backup database
>>> Center 2 - database with a backup database (same as center 1)
>
> Hmm, okay. We're now looking at another, simpler, way to do it. The
> same setup but the clients on Center 1/2 will connect directly to the
> Main center (2Mb leased line). The databases on Center 1/2 will then
> just be a replica of the Main databases.
> The biggest issue though is what happens when the lines go down:
> 1. connect to local database server
> 2. line comes up => changes to local database sent to main database
> 3. connect to main database server
>
> From what I understand this will be possible but I may need to roll
> my own replication code to handle data conflict issues? Especially
> since center 1/2 may be down at the same time and then might change
> the same data.
If you're considering the major choices in replication systems
(Slony-I, Mammoth Replicator), there is no problem with this, as
updates go to the main database server, and there are therefore
necessarily no conflicts.
Both of those systems are asynchronous single master replication
systems.
--
wm(X,Y):-write(X),write('@'),write(Y). wm('cbbrowne','ntlug.org').
http://linuxdatabases.info/info/slony.html
"They laughed at Columbus, they laughed at Fulton, they laughed at the
Wright brothers. But they also laughed at Bozo the Clown."
-- Carl Sagan