Thread: Sync Data

I'm trying to centralize data in a unique db, like it:


---        ---        ---      |DB |           |DB |           |DB |---        ---        ---         \              |
         /           \            |         /      \      -----------    /       |         |/               | "Big" DB
|               |    with    |               | centralized|               | data       |                ------------
 

Information:
1) The "small" DBs sync data with the "big" DB.2) We have 4 linux box above.3) The big DB has a copy of all data. Just
the"small"    DBs inserts data in the "big" one.
 


The question:  Does Postgres do it?

[]'s


Paulo Angelo

This is my first post/reply on this list, but I have been listening for a 
while now (I mostly read the replication ones ;-).  I am interested in 
what developers/users are looking for in a replication/sync solution in 
postgresql, and contributing to that effort.
>I'm trying to centralize data in a unique db, like it:\
<snip, cool ascii art removed >
>Information:
>    1) The "small" DBs sync data with the "big" DB.
What is the connection between small and big Dbs? (LAN/WAN) 
Is there a consistent connection between systems?
>    3) The big DB has a copy of all data. Just the "small" DBs inserts data in the "big" one.
Are you looking for synchronous (before the insert commits) or 
asynchronous (after the insert commits)?
Is there any chance for insert/update conflicts (insert/update same 
record in same table) on the "big" DB?

>The question:  Does Postgres do it?
Short answer, I don't think the current postgresql version will 
accomplish your needs, but there are some postgresql replication projects 
that might get you close depending on the answers to my questions above.  
You can find a list of the projects here.  
http://www.greatbridge.org/genpage?replication_top

Darren Johnson
Source Software Institute

Helo,

On Thu, 31 May 2001 djohnson@sourcesoft.org wrote:

Questions:

>     1) What is the connection between small and big Dbs? (LAN/WAN)


>    2) Is there a consistent connection between systems?

>    3) Are you looking for synchronous (before the insert commits)
>       or asynchronous (after the insert commits)?

>    4) Is there any chance for insert/update conflicts
>       (insert/update same record in same table)
>        on the "big" DB?


-> problem :  I have four "small" DBs that would sync data thru      Internet during the night (128 kbs/s), there is no
    conflicts and there is nobody inserting data in      "small" DBs while its syncing.
 
I think it answers all the questions.


Thanks.


>
> Darren Johnson
> Source Software Institute

Hi Paulo,

It is unclear exactly what you want PostgreSQL to do but it seems to me
that if you want to Sync data into a central database at some point or
other during the day, you should design your application to take care of
it, not rely on PostgreSQL to do anything special.

That is, simply move all the data from the smaller databases to the
central database with a scheduled processing system. 

Gavin

On Wed, 30 May 2001, Paulo Angelo wrote:

> I'm trying to centralize data in a unique db, like it:
> 
> 
> 
>     ---        ---        ---
>        |DB |           |DB |           |DB |
>     ---        ---        ---
>           \              |            /
>             \            |         /
>           \      -----------    /
>            |         |/
>                 | "Big" DB   |
>                 |    with    |
>                 | centralized|
>                 | data       |
>                  ------------
> 
> Information:
> 
>     1) The "small" DBs sync data with the "big" DB.
>     2) We have 4 linux box above.
>     3) The big DB has a copy of all data. Just the "small"
>         DBs inserts data in the "big" one.
> 
> 
> The question:  Does Postgres do it?
> 
> []'s
> 
> 
> Paulo Angelo
> 
> 
> ---------------------------(end of broadcast)---------------------------
> TIP 2: you can get off all lists at once with the unregister command
>     (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)
>