I use Goldengate on classic capture mode to replicate several RAC One Node databases to an external system.
Since the first time there was a relocation of the RAC One Node database, there are two redo log threads (because the DB was temporarily a 2-node RAC).
Now the current thread# has changed and GoldenGate complaints that he cannot access anymore to the redologs.
What was necessary to do was:
1. Check which is the current thread:
SQL> select thread# from v$log where status=’CURRENT’; –> current=1
3. Recreate the extract with:
delete extract e_xx
add extract e_xx tranlog threads 2 begin now
add exttrail /oraacfs/goldengate/gg/dirdat/e_xx/xx extract e_xx
4. Check with
info e_xx
that the RBA for one of the threads increases with the time.
All this should not be needed when running in integrated capture mode.