Replicate using Oracle source fails with missing redo log

[lyka]

Environment

• Qlik Replicate 

If a Replicate task fails with the following error:

00008696: 2021-01-20T10:06:02:964052 [SOURCE_CAPTURE ]E: Cannot find archived REDO log with sequence '7204', thread '1' (oracdc_reader.c:412)

OR

00005624: 2020-06-02T17:05:23:665747 [SOURCE_CAPTURE ]E: Oracle CDC stopped [1022301] (oracdc_merger.c:1178)

Replicate is always reading V$archived_log to get the archived REDO information. The redo log needs to have an active/available status from the V$archived_log

Cause 

This means that Replicate is complaining that it cannot find the redo archive file.

Either

1. the Replicate Oracle Source Endpoint should be using a different Oracle DEST_ID;
2. the file does not exist on disk; or
3. the file exists but that the internal Oracle metadata in v$archived_log view says that it's not available.

The simplest explanation is that the redo log could just have been deleted before Replicate was able to process it.

Resolution

We recommend to at least keep it for 72 hours in case of unexpected weekend issues.

What you can do as well if you encounter an issue like this, is to restore the oracle redo log and then resume the task.

If the redo log cannot be restored, then you need to reload the entire task.

[Steve_Nguyen]

Another issue could happen :

archive logs had to be restored on the source Oracle database server to enable a longer-term replay of change data capture. 

- Task error on similar error : Cannot find archived REDO log with sequence 

Resolution:

• In this recovery scenario, we had get archive logs from the Oracle standby database (this is a DataGuard environment).
• Upon restoring those archive logs to the primary database, they were not cataloged and thus would not populate in v$archive_log, in which Replicate will not read the logs if they aren't present in that view.
• We had to execute a command via Oracle RMAN to re-catalog the files (example)

  • catalog start with '/folder/Oracle SID/folder/OracleSID1_123456';

• Once we did this, Replicate was able to gain acquisition on the restored logs.

[qlik_admin_lm]

• Upon restoring those archive logs to the primary database, they were not cataloged and thus would not populate in v$archive_log, in which Replicate will not read the logs if they aren't present in that view.
• We had to execute a command via Oracle RMAN to re-catalog the files (example)

  • catalog start with '/folder/Oracle SID/folder/OracleSID1_123456';

--- When the Source is Oracle AWS RDS, then what should be done?

[SwathiPulagam]

Hi @qlik_admin_lm ,

As we know there is no way to re-catalog the files in AWS RDS. Please reach out to the AWS team and ask if there is any workaround to achieve this solution.

Thanks,

Swathi