Re: [GENERAL] [Q] Cluster design for geographically separated dbs

2009-03-08 Thread Scott Marlowe
On Sun, Mar 8, 2009 at 1:01 PM, V S P wrote: > Thank you, > Is there a way, in the same  idea, > to make postgresql 'skip' say every 100 numbers when generating > a 'next' in bigserial? > (or to insure that every number generated is evenly divisible by 100, > and then another db would be 99 and so

Re: [GENERAL] [Q] Cluster design for geographically separated dbs

2009-03-08 Thread V S P
Thank you, Is there a way, in the same idea, to make postgresql 'skip' say every 100 numbers when generating a 'next' in bigserial? (or to insure that every number generated is evenly divisible by 100, and then another db would be 99 and so on) In oracle, if I remember right, there was something

Re: [GENERAL] [Q] Cluster design for geographically separated dbs

2009-03-08 Thread Scott Marlowe
On Sat, Mar 7, 2009 at 2:03 PM, V S P wrote: > And wanted to ask what would be the main challenges I am facing with -- > from the experience of the users on this list. > > Especially I am not sure how to for example manage 'overlapping unique > IDs' data. I'm not expert on a lot of what you're d

[GENERAL] [Q] Cluster design for geographically separated dbs

2009-03-07 Thread V S P
Hello, I am designing a db to hold often changed user data. I just wanted to write down what I am thinking and ask people on the list to comment if they have any experiences in that area. My thought is to have say west-cost and east-cost data center and each user will go to either East Coast or