Re: Creating a unique index ID
| From: | Jochem Maas | Date: | Mon, 13 Mar 2006 15:49:34 +0000 |
| Subject: | Re: Creating a unique index ID | ||
| References: | 1 | Groups: | php.general |
| Request: | Send a blank email to php-general+get-231858@lists.php.net to get a copy of this message | ||
Jeff wrote:
isn't that called synchronization? not that it matters. :-)-----Original Message----- From: Jared Williams [mailto:jared.williams1@ntlworld.com] Sent: Monday, March 13, 2006 10:16 To: 'Jeff'; php-general@lists.php.net Subject: RE: [PHP] Creating a unique index IDHmm, actually I'm using circular replication, so it's a bit different than one way replication. A replicates to B which replicates to A.Hey all, I've got a project where I'm taking form information from the user and writing records to several tables in a MySQL database. The problem I'm having is I need to write a unique number for the ID column of the records. Auto increment won't work because I could have conflicts due to replication of the database servers. Anyone have any techique they use for creating unique ID field entries in a db table?Autoincrements can work in a replicated enviroment, lookup mysql settings auto_increment_increment & auto_increment_offset. Each server gets its own unique auto_increment_offset, and auto_increment_increment is set to the number of servers you have. Jared
I've looked at the MySQL docs and can't find anything on how to make Auto_Increment work in Circular Replication.just a thought(tm): it's probably considered hackish (don't know why) but how about setting up a seperate server/instance-of MySQL whose ONLY job it is to provide new ids, and then configure each system to connect [remotely] to the aforementioned db whenever a new id is required... which will probably require some rewriting if you normally use mysql_insert_id() after a query to determine an id (with my proposal you are forced to explicitly generate an id before hand)