How to trace memory resize ops

From: Ingrid Voigt <>
Date: Mon, 26 Mar 2018 12:43:28 +0200
Message-ID: <trinity-f9098d03-8149-47ed-9d2b-149749acdc5b-1522061008634_at_3c-app-gmx-bs25>

Oracle EE on Windows 64 Bit
We've been having problems with memory resize operations shrinking the buffer cache
below specified limits:
Specified:   16GB db_cache_size, 3GB shared_pool_size, 20GB sga_max_size.
Changes to:   approx. 3.5GB db_cache_size, 16GB shared_pool_size
This occurs about 2-3 times per month, usually after hours, severely decreases database
performance and ends with a bounce of the database.
According to MOS 1269139.1 the resizes are expected behavior.
Ist there some way to trace / find out what exactly causes this?
(SQL, background process, other details of the root cause)
My customer is not willing to set _MEMORY_IMM_MODE_WITHOUT_AUTOSGA=false
and risk ORA-04031. If that changes, could one see something with a trace of
Thanks and best regards
Ingrid Voigt
-- Received on Mon Mar 26 2018 - 12:43:28 CEST

Original text of this message