When a plog is corrupted and it is required to recreate the corrupted plog we have to rest reset mine to start parsing again from old redo/archivelog and generate plogs again all over plog. This will re-create all the plogs from the corrupted plog.
Example : Plog 1056
1. First of all shutdown mine and apply.
2. Get REDO LOG sequence from with which plog 1056 was created. Run following command on MINE server.
...
3. Delete records from DBRSAPPLY_PLOG_HISTORY on APPLY server. Copy the table in to some temp table as backup.
...
4. Delete records from DBRSAPPLY_REDO_LOG_HISTORY on APPLY server. Copy the table in to some temp table as backup.
Here SEQUENCE column denotes redo seqnencesequence. In RAC source, a plog is created from multiple redos redo's of different threads, therefore we have to delete records for all threads. ( 2 threads for 2 RAC source).
...
| Note |
|---|
When fetcher is used, delete you don't have to bring down the fetcher and you need not delete archives from mine_stage. Output of step 2 will help you in identifying archives for deletionThe mine would automatically pick up the first archivelog from the mine_stage directory to generate the required PLOG and then move forward. |