Database Software

Oracle 10.2.0.4 on RHEL. In February, we followed Note:805438.1 to open the standby for read-write for a test (after creating a restore point). A few minutes later, we did flashback. It ran fine ever since, receiving and applying archivelogs. Recently we noticed it stopped applying redo and this started in late June. Then I followed Note:836986.1 to roll forward the standby with incremental backup from the primary. During this process, it was found that the standby was in incarnation 2:

RMAN> list incarnation;

using target database control file instead of recovery catalog

List of Database Incarnations
DB Key Inc Key DB Name DB ID STATUS Reset SCN Reset Time
------- ------- -------- ----------- --- ----------- ----------
1 1 ORACP2 4102395133 PARENT 1 20080407 14:13:17
2 2 ORACP2 4102395133 CURRENT 44328058841 20110223 08:33:18

The solution is to set it to 1, because the primary is (and always was) 1:

reset database to incarnation 1;

The DBA team don't recall making any significant change to the standby since the February test and alert.log doesn't show anything relevant. Will "open standby read-write then flashback" leave the standby in the new incarnation, instead of setting it back to the original incarnation? We checked another database which has a standby we did the same test on. The standby is back to the original incarnation as the primary is. So there must be something special about this database.

Secondly, if the standby stayed in a newer incarnation, how could it continue to apply redo received from the primary in its original incarnation?