How to handle OGG processes during DataGuard Switch-Over (Doc ID 1514700.1)

In this Document

Goal

Solution

---

APPLIES TO:

Oracle GoldenGate - Version 11.2.1.0.3 to 11.2.1.0.4 [Release 11.2]
Information in this document applies to any platform.

GOAL

How to handle OGG processes during DataGuard Switch-Over?
 

SOLUTION

Please follow the below steps (first in your development environment) and then in Production,

Steps which need to be followed while performing switch - over to an active standby database
----------------------------------------------------------------------------------------------
Primary Process names: E1, P1 and R1
Secondary Process names: E2, P2 and R2

At this point, E2, P2 and R2 processes are not running.
  
On the Primary production site (Site PDB1)

1.1 Bring down the application hitting the PDB1 database.

1.2 Wait for the open transactions in the queue which are being captured by the extract process.

  send , showtrans
  
1.3 Wait for the pump to read all the latest trails and write to the target trails.

  The current checkpoint for the extract and pump can be monitored through the "info , showch" command. 

1.4 The extract and pump should not be generating any new trails now.

On the Active standby site (Site SDB2)

2.1 Wait for the replicat to process and apply all the pending transactions. 

send , status

Expected output: 0 records to be processed, at EOF

Stop E1, P1 and R1 processes.
 
3.1 Perform the switch over activity (PDB1 will be read-only and SDB2 will be RW now)

4.1 Start the extract E2 using below steps,

alter extract E2, begin now
alter extract E2, etrollover

capture "info E2, showch" output

start extract E2

 
4.2 Start the pump extract P2 using below steps,

alter extract P2, extseqno , extrba 0
alter extract P2, etrollover

capture "info P2, showch" output

start extract P2

 
Monitor the trail file locations to see that extract and pump are writing new trails.

4.3 Start the replicat R2 using below steps, 

alter replicat R2, extseqno , extrba 0
start replicat R2

 

Please note that in this case we cannot directly refer the archive log sequence number or SCN 
since both databases have their own archive log sequence number and different SCNs