• Article Type: General
• Product: Aleph
• Product Version: 18.01

Description:
We are currently running the cold backup weekly on Wednesday mornings on our test server. This Wednesday (7th March) it failed. The Oracle alert log in /exlibris/app/oracle/admin/aleph1/bdump/alert_aleph1.log has the following report:

Stopping background process MMON
Wed Mar 7 01:00:44 2007
Shutting down instance (immediate)
License high water mark = 64
Wed Mar 7 01:05:46 2007
Active call for process 1490 user 'aleph' program 'oracle@abcu (S001)'
SHUTDOWN: waiting for active calls to complete.
Wed Mar 7 01:20:30 2007
MMNL absent for 1204 secs; Foregrounds taking over
Wed Mar 7 02:01:19 2007
SHUTDOWN: Active sessions prevent database close operation
Wed Mar 7 02:01:20 2007
Starting ORACLE instance (restrict)

It looks as though some processes are staying open. We have made the changes to aleph_shutdown as shown in service pack change 628. If you look at the detailed log for the backup, /exlibris/backup/logs/ora_cold_a1_Detail_070307_0100.log you will see that the loop did not reach the limit and so it continued to the section shutting down JBOSS, but nevertheless there appear to be some processes connecting to the Oracle database and keeping it open. Do you know what might be causing this.

Resolution:
The following Oracle Enterprise Manager agent process is causing the Shutdown to hang:
oracle 23290 16094 0 13:04:24 ? 1:04 /exlibris/app/oracle/product/102/jdk/bin/java -server -Xmx256M -XX:MaxPermSize=
oracle 16380 16094 0 Feb 23 ? 22:41 /exlibris/app/oracle/product/102/bin/emagent

Please ensure to shutdown the Enterprise Manager before performing the cold backup.

This seems to have worked. The exact command we used, as root, was

su - oracle10 -c "emctl stop dbconsole > /dev/null"

Our Oracle DBA says that the process which was running, the Oracle Enterprise Manager dbconsole, is something which he sometimes uses to monitor the database. It isn't needed outside his working hours.

• Article last edited: 10/8/2013