vBulletin Search Engine Optimization
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| We have a master/master environment that has tables in innodb. We want to setup a slave that will be used for reporting and some other stuff as well. We want to use MyISAM as the engine on the slave'd server. Since all of the table creations are also part of the replication, is it possible to override the table creates and force them to use a different engine? I have read a few articles on implementing the blackhole engine (for intermediate replication) which would be useful for us when we setup the replication to multiple sites, but this leads to the same question of how to change the engine (which isn't explained in the sample articles I've read). Any help would be greatly appreciated. Gary Wayne Smith |
| |||
| Hi Gary, On Dec 6, 2007 10:56 AM, Gary W. Smith <gary@primeexalia.com> wrote: > We have a master/master environment that has tables in innodb. We want to setup a slave that will be used for reporting and some other stuff as well. We want to use MyISAM as the engine on the slave'd server. > > Since all of the table creations are also part of the replication, is it possible to override the table creates and force them to use a different engine? I have read a few articles on implementing the blackhole engine (for intermediate replication) which would be useful for us when we setup the replication to multiple sites, but this leads to the same question of how to change the engine (which isn't explained in the sample articles I've read). You can set the default storage engine on each of the servers and then don't declare it explicitly in any CREATE TABLE statements. |
| ||||
| > You can set the default storage engine on each of the servers and then > don't declare it explicitly in any CREATE TABLE statements. This seems like the most viable option. Since almost all of the remote tables are created with INNODB it should work fine. I do have one table that isn't but we will convert that before we slave the data to this box. The big problem is test existing dataset is more than 10gb (across several tables/databases). So other suggestions to just recreate the tables by not specifying the engine type isn't particle at this time (but if done at initial design time, we would have been fine). |
| Thread Tools | |
| Display Modes | |
|
|