| what causes the error 1089???!!! [message #57377] |
Tue, 10 June 2003 10:52  |
sheela
Messages: 66
Registered: March 2002
|
Member |
|
|
One of my oracle databases occasionally fails to shutdown completely for an offline backup with this error. After issuing a shutdown immediate ,its somewhere half way I guess and doesn't allow any other process to connect to it. And this remains as such for hours and hours without letting any proces to connect to it. When you try to connect all you get is "Error 1089 ..shutdown in progress". Lets say SID = "AAA". In the trace file (AAA_snp3_21833.trc) after the offline backup was started, I found the following:
/opt/Oracle_8.1.7/rdbms/log/AAA_snp3_21833.trc
Oracle8i Enterprise Edition Release 8.1.7.0.0 - Production
With the Partitioning option
JServer Release 8.1.7.0.0 - Production
ORACLE_HOME = /opt/Oracle_8.1.7
System name: SunOS
Node name: server1
Release: 5.8
Version: Generic_108528-19
Machine: sun4u
Instance name: AAA
Redo thread mounted by this instance: 1
Oracle process number: 11
Unix process pid: 21833, image: oracle@moria (SNP3)
*** 2003-06-07 03:00:07.838
*** SESSION ID:(49.51149) 2003-06-07 03:00:07.806
error 1089 detected in background process
OPIRIP: Uncaught error 447. Error stack:
ORA-00447: fatal error in background process
ORA-01089: immediate shutdown in progress - no operations are permitted
ORA-06512: at "SYS.DBMS_IJOB", line 255
ORA-06512: at line 1
---------------
WHAT CAUSES THIS STRANGE ERROR? WHY DOESN'T ORACLE SHUTDOWN PROPERLY WHEN ISSUED A "shutdown immediate" ? How do I go about debugging this?
Thanks
|
|
|
|
| Re: what causes the error 1089???!!! [message #57387 is a reply to message #57377] |
Tue, 10 June 2003 22:02  |
Devashish Rughwani
Messages: 6
Registered: November 2002
|
Junior Member |
|
|
Problem Description:
====================
You issue a shutdown immediate and receive the following error:
ORA-01089: immediate shutdown in progress - no operations are permitted
Cause: The SHUTDOWN IMMEDIATE command was used to shut down a running
Oracle instance, terminating any active operations.
Action: Wait for the instance to be restarted or contact the database
administrator.
You may be running a shutdown script that performs a shutdown abort, then a
startup, followed by a shutdown immediate. The error is recorded in the alert
log, and a trace file is generated by one or more snp processes.
Solution Description:
=====================
The error is actually expected behavior based on the conditions present during
the shutdown immediate. The following workarounds are available:
Workaround 1:
=============
Modify your shutdown script to issue a shutdown normal instead of a shutdown
immediate.
Issuing a shutdown normal allows the snp processes to complete their work
prior to shutting down, thus avoiding the error. The implication of this
workaround is that the database will not shut down until all user and snp
processes have completed.
Workaround 2:
=============
Increase the JOB_QUEUE_INTERVAL parameter in the INIT.ORA file.
*NOTE: Increase by increments and test to determine what interval is
necessary to avoid the error.
This increases the time before the snp processes wake up after starting
up the database, allowing the shutdown immediate to be issued before the
snp processes wake up.
Workaround 3:
=============
Start the database with a different INIT.ORA file. Modify the startup
script to use a separate INIT.ORA file that does not include the
JOB_QUEUE_PROCESSES or JOB_QUEUE_INTERVAL parameters.
For Oracle 8i versions you can set the JOB_QUEUE_PROCESSES to 0, this will
ensure that no snp processes startup.
Since no snp processes will start during the startup, no error will occur
during the shutdown immediate.
Explanation:
============
During the startup portion of the shutdown script, Oracle starts a number of
snp processes based on the "init.ora" parameter "JOB_QUEUE_PROCESSES".
Depending on how long Oracle takes to process the shutdown immediate portion of
the shutdown script and the value specified in the "init.ora" parameter
"JOB_QUEUE_INTERVAL", one or more of the snp processes may wake up to check the
jobs queue. The smaller the value of "JOB_QUEUE_INTERVAL", the sooner the snp
processes will wake up. If these processes wake up between the time the
database was started and the shutdown immediate was issued, the ORA-1089 error
will occur. The snp process cannot continue because a shutdown is in process.
|
|
|
|
| Re: what causes the error 1089???!!! [message #57395 is a reply to message #57377] |
Wed, 11 June 2003 16:09 |
pwl
Messages: 22
Registered: May 2000
|
Junior Member |
|
|
Maybe slightly off the topic - but the problem is
the shutdown immediate. Do not do it otherwise you
will wait until Oracle believes that there are no
connections - and if it thinks there are, but they do
not really exist you will wait forever.
The best plan is to use a script that works :-
1. Alter system checkpoint;
2. Alter system switch logfile;
3. shutdown abort;
The steps are as follows:
1 forces a checkpoint so all data is written to disk
2.allows the archiver to write the logfile
3. oracle stops
On startup there will be a pause for recovery, but
only what is in the latest logfile.
The benefits - full recovery and no waiting to
shutdown.
cheers
pwl
|
|
|
|
| Re: what causes the error 1089???!!! [message #57406 is a reply to message #57377] |
Thu, 12 June 2003 14:13 |
Trifon Anguelov
Messages: 514
Registered: June 2002
|
Senior Member |
|
|
Nothing wrong with the fact that no connections are allowed when you issued "SHUTDOWN IMMEDIATE"; The db instance has to be closed, after this command, not open and allowing connections.
The problems is why it takes "hours and hours" to shutdown. Some of the reasons and the possible fixes are listed Here
Hope that helps,
clio_usa
OCP - DBA
Visit our Web site
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
