Distributed
polling mechanism is used when a DB poller process is deployed in an
active/active cluster environment. In that case a deployed process will have
two instances of process polling on the same table. Distributed polling
techniques can be used for inbound Database Adapters to ensure that the same
data is not retrieved more than once.
For
a typical DB adapter polling from oracle
database table, it is very easy to add distributed polling mechanism by
just checking Distributed polling check box in the adapter creation wizard.
But
when a DB adapter is polling from a sybase database table, the same
configuration will not work and we might face issue like the same record is
polled twice in an active-active cluster environment. To work distributed
polling for a sybase database poller properly we will need to add few
additional steps as part of the adapter configuration. Below are the steps to
apply distributed polling for a sybase database adapter…
1. Before configuring the
database adapter go to the connection factory settings from console->Deployments->DbAdapter->Configuration->Outbound
Connection Pool.
Click on your connection factory which is using the Sybase data
source and change the platformClassName value
to org.eclipse.persistence.platform.database.DatabasePlatform.
Save and active changes.
2.
Now go to the adapter configuration wizard. In the page where we
select the column name and read/unread value set Resereved value as R${weblogic.Name-2}-${IP-2}-${instance}
1.
Go to next page and check distributed
polling check box. We will see a new pre poling SQL is generated this time.
Click next and finish the adapter configuration wizard…
Now
the polling process can be deployed and tested in a cluster environment.
Note:
The read column (RequestStatusCd in above example) should have the column
length to take reserved value (R${weblogic.Name-2}-${IP-2}-${instance})
which may vary from one environment to another.
