My Oracle Support Banner

Why the Parallel replicat have high memory consumption (Doc ID 2811472.1)

Last updated on JULY 20, 2024

Applies to:

Oracle GoldenGate - Version 19.1.0.0.0 and later
Information in this document applies to any platform.

Symptoms

When parallel replicat processes are been started, the node crashed with out of memory
and CPU (high load scenario where all the CPU and memory resources got utilized)

It was also noticed that, the replicat process id which is showing in the report file (Mapper ID process id's) are consuming
more memory around 80GB.

If the workload has a lot of lobs then the cachemgr will be used by the
Mappers and Appliers. Please try setting "cachemgr cachesize 2G". The
default is 16GB (with a hard limit of 24GB). Note that cachmgr stats in the
PR report file are not valid as they only reflect the cachemgr use by the
main replicat process which does not actually apply the transactions. Also set the
"unifiedreport false" to observe on a clean shutdown the cachemgr values for the Appliers and Mappers
by looking in their respective report files. In particular "vm used max" would be interesting to know.



Changes

Tested parallel replicat with the following settings:
cachemgr cachesize 2G
unifiedreport false

Cause

To view full details, sign in with your My Oracle Support account.

Don't have a My Oracle Support account? Click to get started!


In this Document
Symptoms
Changes
Cause
Solution
References


My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.