We’re facing a strange problem that we’ve never encountered before.
We’re having a Dell PowerEdge R330 Windows server with 8 GB RAM and Oracle 11g SE1 v.11.2.0.4.
The system has 3 databases configured and is definitely not overloaded (on the contrary).
Observation:
- Approx. 6.5 days after a system restart the alert_<db>.log’s starts logging the following:
ORA-25253 encountered when generating server alert SMG-3503
- E.g.:
Thu Nov 01 16:47:58 2018 Starting ORACLE instance (normal)
Wed Nov 07 06:33:13 2018 ORA-25253 encountered when generating server alert SMG-3503
…
Tue Jan 15 11:08:04 2019 Starting ORACLE instance (normal)
Mon Jan 21 00:31:35 2019 ORA-25253 encountered when generating server alert SMG-3503
…
Mon Jan 21 10:17:40 2019 Starting ORACLE instance (normal)
Sun Jan 27 00:12:11 2019 ORA-25253 encountered when generating server alert SMG-3503
The above happens on all 3 databases more or less simultaneously, which indicates that the problem may somehow be related to the listener.
Some time after the ORA-25253 entry the listener also starts complaining with entries like:
29-JAN-2019 06:44:16 TNS-12531: TNS:cannot allocate memory
We’ve tried reinstalling Oracle from scratch, but the problem reappeared.
The system is more or less running in idle mode.
There is no indication of “someone” spawning an abundance of processes or something similar.
Any suggestions / advice?
BR Per