-
Title
Seeing data going back to primary after upgrade -
Description
After upgrade from 5.1.x to 5.3.3, seeing data being replicated back from secondary to primary. -
Cause
Due to post running and capture not current on redo log before shareplex shutdown on secondary in H/A or Peer to Peer environment. Shareplex_trans table gets recreated after shareplex upgrade (running ora_setup) causing object_id to change. -
Resolution
Please follow this procedure if you have H/A or Peer to Peer set up where you have a reverse replication going from secondary to primary.
You can follow this procedure for the upgrade.
1. On Primary:
sp_ctrl> stop post
<if></if>sp_ctrl> flush o.SID
2. On Secondary:
a. Shutdown application or stop user access if this is peer to peer. there should be no user access except shareplex post if this is H/A.
b. Check and make sure post is stopped due to flush.
c. Check and make sure capture is current with oracle
sp_ctrl> show capture detail
You can do a log switch if you wish and verify capture is on the current redo log.
d. Shutdown shareplex on both servers and do the upgrade.
After upgrade, check post queue on Primary you should not see any messages.
Note: Restart post on primary if this is peer to peer and allow user access on secondary.
-
Defect ID
ST29785