ORA-01341: LogMiner out-of-memory

The Problem

Having recently set-up and configured an environment using Change Data Capture (CDC), we were running through a test which was a representative workload on the main system. After some time we encountered the following error:

ORA-01280: Fatal LogMiner Error.
ORA-01341: LogMiner out-of-memory

This error was at the downstream database. We were using Asynchronous CDC Autolog Archive. My first attemp to fix this error was to increase the STREAMS_POOL_SIZE. This didn’t help and again we found ourselves back to the same error.

The Solution

It is possible to increase the amount of physical memory assigned to each capture process, as long as you have enough physical memory in your database you can increase the parameter _SGA_SIZE for the capture causing the problem using the following:

exec dbms_capture_adm.stop_capture(‘CAPTURE_NAME’);
exec dbms_capture_adm.set_parameter(‘CAPTURE_NAME’,’_SGA_SIZE’,’50’);
exec dbms_capture_adm.start_capture(‘CAPTURE_NAME’);

You can read more about this in Metalink note 336705.4

If you are running into the error ORA-01372: Insufficient processes for specified LogMiner operation, please read my other post. I’ve also come across the ORA-31514: change set %s disabled due to capture error error message.

Like it, share it...

Category: LogMiner

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *