Check if you have JOB_QUEUE_PROCESSES 0
and the schedules for jobs in DBA_JOBS.
A running job will prevent a shutdown.
Hemant
--- DENNIS WILLIAMS [EMAIL PROTECTED] wrote:
Jeremiah, Gaja
Thanks for the tips. My theory is that since 3 production
instances were
affected, it may be easy to
1. Turn on PMON and SMON tracing for your instance and look at their
trace files during shutdown
2. Look for segments of type TEMPORARY with many extents that need
deallocating prior to shutdown - this will be done lazily during
shutdwn immediate
3. Look for large transactions that need
Dennis,
Just to add to Jeremiah's list:
I have also seen the aforementioned symptom, when
there were SQL*Plus sessions that are out there on
the login screen with no takers.
Cheers,
Gaja
--- Jeremiah Wilton [EMAIL PROTECTED] wrote:
1. Turn on PMON and SMON tracing for your instance
and
And, if you have any job that is waiting on a DBMS_PIPE.
-Original Message-
Krishna Vaidyanatha
Sent: Sunday, March 23, 2003 3:09 PM
To: Multiple recipients of list ORACLE-L
Dennis,
Just to add to Jeremiah's list:
I have also seen the aforementioned symptom, when
there were SQL*Plus
Jeremiah, Gaja
Thanks for the tips. My theory is that since 3 production instances were
affected, it may be easy to find.
Dennis Williams
DBA, 40%OCP, 100% DBA
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original Message-
Sent: Sunday, March 23, 2003 11:54 AM
To: Multiple recipients of list
DENNIS WILLIAMS wrote:
We have been running Oracle 8.1.6 unchanged for several years. Within the
past 3 days, our cold backup scripts have had a shutdown immediate hang. In
the alert log the message is:
Shutting down instance (immediate)
Sun Mar 23 00:09:10 2003
SHUTDOWN: waiting for active
Dave,
Worth Checking -
Sometimes the OracleSIDAgent service or the OracleSIDDataGatherer
service, if running, will cause the SHUTDOWN IMMEDIATE to hang. Try
shutting such services down by including a 'net stop OracleSIDAgent..'
etc. in your script, just before you issue the SHUTDOWN
Dave:
The message in the alert log means that there are uncommitted transactions.
During a shutdown immediate uncommitted transactions are rolled back and
Oracle must wait until this rollback is complete before the db can be
actually shut down.
You could do a shutdown abort, but then you would
Dave,
We've periodically had this problem with various jobs that
were running at the time of the shutdown immediate. I try
to isolate the cause by killing the various processes I still
see running via ps -ef | grep SID.
Sometimes the problem has been our RMAN backup that
is running,