Deployments performed in switch mode require two databases, where the repository assets are updated on the target site’s inactive database while its active database runs undisturbed. These databases should begin with identical content. To accomplish this, you can use the tools provided by your database vendor to make the second database identical to the first database. Alternatively, after you configure the target for switch deployments, you can perform a full deployment to the target.
When configuring target site repositories for switch deployments, make sure that their definition files do not contain operation tags such as <additem>
, <updateitem>
, and <removeitem>
. These tags cause switch deployments to fail.
To configure target repositories for switch deployments, perform the following procedure on all ATG servers that will switch their underlying active and inactive database:
Configure data sources that connect to the two databases.
Configure a
SwitchingDataSource
to switch between the two underlying data sources.Configure the repository components to point to the
SwitchingDataSource
. Also, set each repository component’scacheSwitchHot
property to true.Add the
SwitchingDataSource
to the agent’s list of switchable data stores inDeploymentAgent.switchableDataStores
.To do so, layer on a
DeploymentAgent.properties
, placing the new configuration file in yourlocalconfig
directory at/atg/epub/
. Make sure to specify theSwitchingDataSource
by its fully qualified Nucleus path.
Note: For detailed information on steps 1-3, see the ATG Installation and Configuration Guide.
Configuring a New ATG Server for Switch Deployment
When adding an ATG server to an existing target, configure its repositories and SwitchingDataSource
like those on the existing ATG target servers. Verify that:
It has the same underlying data sources.
Its initial data source is the same as their current data source, so the new server’s agent application starts up using the same data source.