The oracle probe appears to have a memory leak. The oracle probe is monitoring database instances remotely (not on the local database server). The memory is continuously increasing.
Environment: UIM v8.31
oracle probe v5.01 GA
OS the probe is running on: RHEL 6, 2.6.32-573.22.1.el6.x86_64
hub v7.80
robot v7.80, where oracle probe sits
12 G of RAM available on the robot where the oracle probe is running
profiles: > 51 on the Linux machine, 55+ connections
size of the oracle.cfg on the Linux machine is ~11MB
Oracle 11g and/or 12c database instances are being monitored
Oracle cluster and/or single instances (probe running on single instance RHEL)
Oracle 12c client has a known memory leak as per Oracle Support. Memory leak is between 110-120k per Oracle client session (connection) when using the oracle probe.
1. Install Oracle client 11.2.0.3 (11g libraries are installed as part of the client install.) The 12c libs were still in place.
2. The Robot environment had to be configured for the 11g client.
For example:
variable name variable value
LD_LIBRARY_PATH /orah4/product/11.2.0.3/lib
NLS_LANG
ORACLE_BASE /orabase
ORACLE_HOME /orah4/product/11.2.0.3
PATH /orah4/product/11.2.0.3/bin:$PATH
3. Cold start the oracle probe (Deactivate-Activate)
The memory leak issue should be averted.
The Oracle SR number is:
Please also refer to the oracle probe Help doc: