If you have 1m users,.

 
Digangi: Is there any help on golden gate – Oracel -mYsql

Schnitker: Danblack: it’s kind of a unique contraint. It is a sometimes-its-not-unique constraint

Sabbatini: Rrk_: what’s your question ?

Karlsen: JesusTheHun: it is simple. A,B is unique. or A,B isn’t unique – maybe A,B,C is sufficinently unique. We can’t remove the contraint for you. its your choice. MySQL doesn’t support somtimes-its-not-unique contraint types. sorry.

Maimone: How can i install Golden gate in mysql

Hubright: Rrk_: i think that’s a question for the golden gate folks.

Buker: Yes .Please let me know if there any other channel for the same

Regalbuto: Looking at the oracle pages you might be looking for an email/phone number of a sales office.

Hintz: I’m wondering how people implement an incremental update “checker”? Is it better to store latest id or something in separate table, and the checker just query this table to know if a specific users has new data or not.

Darmody: Fairuz: start by explaining what you want to do

Cypress: If you are thinking of a message queue, message queue implementations are better at doing message queues than relational databases.

Cashwell: I have a local db in a device. Its data comes from a remote DB. A user will insert new data to this remote DB. So I need to know the incremental update that I need to fetch from the remote DB, to insert to the local DB.

Borrigo: I hope I explain it well

Braune: Fairuz: See http://dev.mysql.com/doc/refman/5.6/en/replication.html

Seymer: Will replication still be valid if I say the device in question is an Android device using SQLLite and the remote DB is in my backend server running MySQL?

Stampe: In addition to that, in the device I only need the data for a specific user

Gunyan: First sentence of the link provided says mysql to mysql

Galbo: While in the remote DB, there are data for all the users

Pucker: You might add a “last update” or “inserted at” timestamp column and remember time of last syncing

Monarrez: Fairuz: jkavalik that’s what most people do

Pareja: That’s I plan to do, just brainstorming maybe a general practice that I’m not aware of it

Stencel: It depends on the amount of data you need to sync

Loshe: My concern is on the frequency and the number of tables involved

Hoesing: Let say I want to sync every minute

Jekot: Fairuz: wrap in in a web API. use http header around expiry and use that to match the underlying impliementation. a trigger to update a users row might be worthwhile.

Riska: So every minute, it means I need to query x number of tables just for one user?

Soos: If it’s just update a row or so, then add a last_update column, if you have more data, maybe perform the few queries since last sync. if you have more data, create a table of sync, and put a trigger on update to set the data to be sync next time for this table

Delliveneri: This might be a message queue problem. – even if some of the underlying data is in the database.

Kemp: Thinking queue per person. don’t know how that scales.

Bullion: Fairuz: can you tell us more about the environement of the project ?

Oshima: The device in question is an Android device using SQLLite and the remote DB is in my backend server running MySQL. The device will communicate with the backend using REST APIs. A user will use a web application to update or insert new data to the backend.

Murri: What are you using the local db for? caching?

Ace: My initial idea is to do a polling each minute to the backend to check for new data

Penepent: Imagine that polling done by 1 million users

Mulgrew: So it’s more a pre-cache

Pestronk: That’s my concern as well

Veras: What if I have 1 million user

Jubie: How people usually tackle this kind of problem

Vibbert: If you have 1m users, you’ll have time and money to think about it