Skip to Main Content
  • Questions
  • shutdown immediate hung forever and ORA-03113

Breadcrumb

May 4th

Question and Answer

Tom Kyte

Thanks for the question, joon.

Asked: June 08, 2001 - 12:49 pm UTC

Last updated: July 16, 2013 - 4:58 pm UTC

Version: 8.1.6.2

Viewed 10K+ times! This question is

You Asked

Hi
Our client using
hp-ux 11.0 and oracle 8.1.6.2.

We have problem in backup.
Backup script doing following sequence.
1.shutdown abort
2.startup
3.shutdown immediate
4. backup all files that belongs to db.

The problem is Onece a month shutdown immediate hung forever.
We found in the morning following message in the alert log.
"SHUTDOWN: waiting for active calls to complete."
But No jobs run before backup and all tablespace pctincrease 0
except system tablespace, temp tablespace temporary.

whenever shutdown hung ,same time ORA-03113 happen.

Please help me.
thanks.









and Tom said...

well, the easiest and best solution to this is to NOT shutdown to do a backup. HOT backups are easy, and do not make you shutdown.

I can only guess at what may have happened. The shutdown abort aborted a really long running transaction that would take a while to roll back. when you started up, we rolled forward, opened the database and started to roll back (in the background -- slowly so as to not overwhelm the system). You did a shutdown immediate -- which does a CLEAN shutdown. It must wait for the rollback to finish. It could take longer to rollback then it did to do the transaction in the first place.

So, your shutdown abort make it so the shutdown immediate would have to wait a really long time.

This is a good reason why you want to do HOT backups -- so you don't run into issues like this.


follow up to comment one below

Shutdown immediate is a clean shutdown. It simply kicks off active transactions before doing a clean shutdown whereas shutdown normal WAITS for them to complete.

A shutdown immediate leaves the database in the same closed state as shutdown normal does.

Rating

  (36 ratings)

Is this answer out of date? If it is, please let us know via a Comment

Comments

How about startup open restrict/shutdown normal

Moorthy Rekapalli, June 12, 2001 - 11:19 am UTC

How about doing the following:

svrmgrl <<EOF
shutdown abort
startup open restrict
shutdown normal
EOF

That way, you will get a normal shutdown before taking offline backups. Shutdown immediate might work just fine for offline backups. But, shutdown normal will give peace of mind.

ora 3113

mo, October 03, 2002 - 6:29 pm UTC

Tom:

We have several instances/databases running. My main database uses database links to select data from others. The problem I am having is that everytime the dba shuts down the instance I get this:

ORA-02068: following severe error from INS_LINK
ORA-03113: end-of-file on communication channel
ORA-06512: at "SYS.DBMS_SYS_SQL", line 1474
ORA-06512: at "SYS.OWA_UTIL", line 1663
ORA-06512: at "INSADMIN.NEW_PO", line 516
ORA-06512: at line 7

He says that he has to drop and re-create the links afterwards. I am thinking something is wrong here. After you start a database the links should re-establish themselves and I should not get this error the first time I call the link.

Do you think this issue is related to "shutdown immediate" or "shutdown". Any hints?

BTW when I searched your site for ORA-03113 I got 68 records found and you only displayed 15.

Thank you,


Tom Kyte
October 03, 2002 - 6:58 pm UTC

You got APPROXIMATELY 68 records found. Just like google does (and hey, if google can do it, so can I cause google is just the coolest)

see
</code> http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:3489618933902 <code>
and search (ctl-f) for google.


Your DBA needs to open a tar with support -- the links should not need to be dropped and recreated, there is something fundementally wrong there.

Sql Report Writer 1.1

Bipin Ganar, July 31, 2003 - 6:51 am UTC

Hi Tom,
We have shifted our oracle 8.1.7 Database From open VMS to Windows 2000 Oracle 9.2.0. My other Applications are working without any problem But Reports which are using report writer 1.1 is giving me err Ora-03113

As follows
SQL*ReportWriter : Version 1.1.14.16.1-Product on Tue Jul 29 14:56:49 2003 Copyright (c) 1987 Oracle Corporation,California,USA
All Rights reserved Wrldwid Oracle Error Number -3113
ORA-03113 : end-of-file on communication channel

Can u help me urgently


Tom Kyte
July 31, 2003 - 7:21 am UTC

umm, contact support -- however, that is such "old stuff" -- i have a feeling it is going to be "sorry, please upgrade" at this point.


if report writer 1.1 was built on 7.3 support libraries - I am 100% sure that will be the answer since 7.3 clients connecting to 9iR2 is known not to work and is not supported at all.

something for Mo

A reader, July 31, 2003 - 10:48 pm UTC

Whats the database link u use linking to i mean whats the source database, si it oracle or some other database, i hv used hetergenous tables and faced similar problems, luckily hv been able to identify and get a workaround on that, let me know if thats the architecture.



ora-03113

atul, March 31, 2004 - 10:19 am UTC

Hi,

When i was running a big update (60,000,00) rows on a table.
I got a error
ORA-03113,
My co-DBA says,this is because PGA is full.

Is that valid reason for mentioned error?if yes how to check
if PGA is full or not?

Thanks & Regards,
Atul

Tom Kyte
March 31, 2004 - 10:40 am UTC

check out the trace file.

pga getting full would return ora-4030

3113 means "something bad has happened", typically necessitates a call to support.

ORA-03113

ruby, April 15, 2004 - 11:20 am UTC

I am getting
ERROR at line 1:
ORA-03113: end-of-file on communication
when executing following statement twice in same session. but it is working well if I close the session and relogin and run ,I need to set this variable to new times using stored procedures after 30 minutes
********
commit;
f:='set endtime ='||v_end_date||' , starttime ='||v_start_date ||', rowcount='||rows;
E:=DBMS_HS_PASSTHROUGH.EXECUTE_IMMEDIATE@IHISTLINK(f);
commit;
********

Tom Kyte
April 15, 2004 - 11:23 am UTC

you'll have to contact support for that. rowcount seems sqlserverish?

Shutdown 2 minutes after restarting database!

A reader, May 04, 2004 - 5:53 pm UTC

Hi Tom,

I am using oracle 9.0.1 on HP. When I startup the database, it will open the database and even I can issue the command like 'select * from tab' and it works, but after 2 minutes, it shutdown automatically! I even restart the server but it didn't fix the problem.

Thanks for your time.


Tom Kyte
May 04, 2004 - 7:40 pm UTC

please contact support (after reading your alert log to see if it is something obvious)


the database won't shutdown automagically -- something else is afoot.

Thanks

A reader, May 05, 2004 - 12:53 pm UTC

Thanks Tom. The system tablespace was corrupted. So after I restored the back it works well.



A reader, September 17, 2004 - 7:28 pm UTC

Tom,

ORA-01090: shutdown in progress - connection is not permitted

I get the above message when I try to connect as sysdba. There is no way I can do a normal shutdown now. Somewhere something is shutting it down and it won't even let me in. Any advice?

Tom Kyte
September 17, 2004 - 8:31 pm UTC

I think you are trying to connect as sysdba wrong maybe? sysdba accounts can log in during a shutdown:

[tkyte@tkyte-pc-isdn tkyte]$ sqlplus /

SQL*Plus: Release 9.2.0.4.0 - Production on Fri Sep 17 20:10:14 2004

Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.

ERROR:
ORA-01090: shutdown in progress - connection is not permitted



[tkyte@tkyte-pc-isdn tkyte]$ sqlplus "/ as sysdba"

SQL*Plus: Release 9.2.0.4.0 - Production on Fri Sep 17 20:10:23 2004

Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.

Connected.
idle>



A reader, September 17, 2004 - 8:58 pm UTC

Tom,

I was trying to connect as "/ as sysdba"

Tom Kyte
September 17, 2004 - 9:10 pm UTC

lets see -- just like I did -- cut and past -- no editing please!

shutdown immediate hangs.... 9.2.0.4

reader, March 14, 2005 - 9:20 am UTC

What could be the possible reasons that shutdown immediate hangs? Is it only because of the time it takes to rollback any active transactions? Is this why it appears to be hanging? Thanks.

Tom Kyte
March 14, 2005 - 9:47 am UTC

shutdown immediate will kill and then rollback open sessions, it could take a very long time if the transactions have been running for a very long time.

shutdown in progress

Jun Chen, April 05, 2005 - 2:40 am UTC

Hi Tom,

I have one session opened to shutdown database (forgot to type immediate before hit enter), and seeing there is user still logged in from unix. I contacted user but they are not there; and I am able to connect as sysdba, however I couldn't do anything - select statement give me 'not logged on ' error message; and if I initiate another 'shutdown immediate', it will give me error 0ra-01090: shutdown in progress - connection is not permitted. What should I do to speed up shutdown process?

Thanks!

June

Tom Kyte
April 05, 2005 - 7:31 am UTC

ctl-c the shutdown if you like and shutdown abort, or just shutdown abort from another session.

Shutdown Immediate Hangs

Tony Johnston, April 05, 2005 - 9:26 am UTC

Hello Tom,

On some of the databases I work on, shutdown immediate seems to just hang there for a very long time, the longest was one hour before I ctl-c it, I then shutdown abort, and startup as usual and all seems OK. Should shutdown immediate hang like this? and is there any danger in constantly using shutdown abort? As I understand it, startup after shutdown abort will apply the redo logs correctly.

Thanks

Tom Kyte
April 05, 2005 - 12:05 pm UTC

shutdown immediate will roll back everything and wait until it is done.

abort just kills it and we do recovery upon restart.



ORA-03113

Sam, November 08, 2005 - 5:17 pm UTC

Tom:

I have a web application that uses a couple of database links to another databases to lookup information. Server is AIX 4.3 and database is 8.1.2.

Sometimes we get
ORA-03113: end of file on communication channel

1. Can this be in any way due to links or application. It seems to me it is NET8 or database configuration or an oracle server bug issue.

2. Did oracle drop support for 8i running on AIX 4.3. DBA says he cant put a TAR due to that?

3. Can this be due to shutting down the app server every week?

4. How do you pinpoint the problem?




Tom Kyte
November 08, 2005 - 10:47 pm UTC

1) yes it could be.

2) dba is only partially correct. 8i is out of ECS (error correction support) but you can still log tars.

3) doubtful.

4) I'd start by working with support - they can help you diagnose this.

ORA-03113

mo, November 09, 2005 - 11:08 am UTC

tom:

1. what do you mean by "error correction support". Does it mean oracle is not releasing any fixes for this version?

2. Can you refer me to a site or a link or anything that says customers can still log tars for 8i database ruinning on aix 4.3.

3. What in an application design could be causing this error?

4. would upgrading to 9i or merging several databases resolve this issue?

Thank you,



Tom Kyte
November 11, 2005 - 10:07 am UTC

1) correct, if the patch does not already exist, no new patches are developed. work arounds are attempted to be found but that is it, the code is pretty much "done"

2) metalink.oracle.com has all of this information.

3) nothing, it could be a bug, it could be a configuration issue, it could be a firewall timeout, .........


4) yes, no, maybe... You have to sort of find the cause, why is the backend process going away - why is the network getting cut off.

A reader, December 22, 2005 - 3:57 pm UTC


shutdown hung

oradba, January 25, 2006 - 6:43 pm UTC

Tom,

I valiated the hot backup and did shutdown abort

Startup
alter system switch logifle;
/
/
/
shutdown immediate

It is taking lot of time


Wed Jan 25 15:26:22 2006
Shutting down instance (immediate)
License high water mark = 16
Wed Jan 25 15:26:25 2006
ALTER DATABASE CLOSE NORMAL
Wed Jan 25 15:31:36 2006
Waiting for smon to disable tx recovery.

it takes around 1 hour

Please suggest what should I do to make it fast?

Tom Kyte
January 25, 2006 - 7:02 pm UTC

shutdown abort is pretty fast.

shutdown immediate will do what it needs to do to shutdown clean - big SGA with lots of changes to checkpoint, it'll take a while.

What is the reason to shutdown?

shutdown hang

oradba, January 25, 2006 - 7:16 pm UTC

Our production backup is coming to backup server and I need to copy datafiles from backup server to development server so before start copy I want to validate the backup.

I notice one thing if I keep database open for 20 minutes then shutdown immediate takes only 2 mins.

Very strange behaviour why?

Tom Kyte
January 25, 2006 - 7:57 pm UTC

hot backup, you don't shutdown just to do a backup.

I don't know what you mean by "if i keep the database open for 20 minutes", why wouldn't the database be open for a *long long time*??? How did the database you keep open for less than 20 minutes get to where it is in the first place?


strange behaviour

A reader, February 21, 2006 - 5:15 am UTC

I have a table t1 with zero rows.
I INSERT INTO t1.... , it was taking along time so after
5 hours I killed the session and then shutdown abort.
Then I startup, it opened the databse.

Then I did SELECT COUNT(*) FROM t1, its taking very
long time, it hadnt returned to me anything till now.

I know that there is a background process going on
for rollback the INSERT i had cancelled.

1- t1 had zero rows before INSERT, the INSERT was killed,
so t1 still has zero rows, then why its taking so long to
count zero rows.

2- Regardless of the above scenario, is there a way
to kill the active rollback transactions? if yes what will
be the disadvantages of it.

Thanks



Tom Kyte
February 21, 2006 - 7:50 am UTC

rolling back is a logical - not physical - operation.

You advanced the high water mark of the table.


In order to rollback a massive (conventional path) insert, Oracle will DELETE the rows you just inserted.


So, you might want to truncate that table.

A reader, February 21, 2006 - 10:41 am UTC

So can I say that if there are zero commited and
millions of uncommited rows in a table then
SELECT COUNT(*) from that table will take a long time,
until the rollbacking of that millions of rows is not
completed?

Tom Kyte
February 22, 2006 - 7:57 am UTC

you can say that, and you can say

if I insert millions of rows and roll them back, it'll take a long time to full scan the table since the INSERT advanced the high water mark of the table and a rollback WILL NOT lower it.  A rollback of an insert is much like a delete!  it is as if we:

insert into t select * from big_table;
delete from t;
commit;

instead of

insert into t select* from big_table;
rollback;


<quote src=Expert Oracle Database Architecture>

It is a common misconception that undo is used to restore the database physically to the way it was before the statement or transaction executed, but this is not so. The database is logically restored to the way it was—any changes are logically undone—but the data structures, the database blocks themselves, may well be different after a rollback. The reason for this lies in the fact that, in any multiuser system, there will be tens or hundreds or thousands of concurrent transactions. One of the primary functions of a database is to mediate concurrent access to its data. The blocks that our transaction modifies are, in general, being modified by many other transactions as well. Therefore, we cannot just put a block back exactly the way it was at the start of our transaction—that could undo someone else’s work! 

For example, suppose our transaction executed an INSERT statement that caused the allocation of a new extent (i.e., it caused the table to grow). Our INSERT would cause us to get a new block, format it for use, and put some data on it. At that point, some other transaction might come along and insert data into this block. If we were to roll back our transaction, obviously we cannot unformat and unallocate this block. Therefore, when Oracle rolls back, it is really doing the logical equivalent of the opposite of what we did in the first place. For every INSERT, Oracle will do a DELETE. For every DELETE, Oracle will do an INSERT. For every UPDATE, Oracle will do an “anti-UPDATE,” or an UPDATE that puts the row back the way it was prior to our modification.

Note    This undo generation is not true for direct path operations, which have the ability to bypass undo generation on the table. WeÂ’ll discuss these in more detail shortly.

How can we see this in action? Perhaps the easiest way is to follow these steps:

    1.    Create an empty table.
    2.    Full scan the table and observe the amount of I/O performed to read it.
    3.    Fill the table with many rows (no commit).
    4.    Roll back that work and undo it.
    5.    Full scan the table a second time and observe the amount of I/O performed.

So, weÂ’ll create an empty table:

ops$tkyte@ORA10G> create table t
  2  as
  3  select *
  4    from all_objects
  5   where 1=0;
Table created.

And then weÂ’ll query it, with AUTOTRACE enabled in SQL*Plus to measure the I/O. 
Note    In this example, we will full scan the tables twice each time. The goal is to only measure the I/O performed the second time in each case. This avoids counting additional I/Os performed by the optimizer during any parsing and optimization that may occur. 

The query initially takes three I/Os to full scan the table:

ops$tkyte@ORA10G> select * from t;
no rows selected
                                       
ops$tkyte@ORA10G> set autotrace traceonly statistics
ops$tkyte@ORA10G> select * from t;
no rows selected
                                      
Statistics
----------------------------------------------------------
          0  recursive calls
          0  db block gets
          3  consistent gets
...
ops$tkyte@ORA10G> set autotrace off

Next, we’ll add lots of data to the table. We’ll make it “grow” but then roll it all back:

ops$tkyte@ORA10G> insert into t select * from all_objects;
48350 rows created.
 
ops$tkyte@ORA10G> rollback;
Rollback complete.

Now, if we query the table again, weÂ’ll discover that is takes considerably more I/Os to read the table this time:

ops$tkyte@ORA10G> select * from t;
no rows selected
                                            
ops$tkyte@ORA10G> set autotrace traceonly statistics
ops$tkyte@ORA10G> select * from t;
no rows selected
                                         
Statistics
----------------------------------------------------------
          0  recursive calls
          0  db block gets
        689  consistent gets
...                                           
ops$tkyte@ORA10G> set autotrace off

The blocks that our INSERT caused to be added under the table’s high-water mark (HWM) are still there—formatted, but empty. Our full scan had to read them to see if they contained any rows. That shows that a rollback is a logical “put the database back the way it was” operation. The database will not be exactly the way it was, just logically the same.
</quote>
 

truncate or drop

A reader, February 23, 2006 - 10:29 am UTC

Ok, but now I am
trying to
truncate table t1;
its giving me resource busy and acquired with nowait specified.

and if I
drop table t1;
its giving me same resource busy and acquired with nowait specified.

Is there a way to kill that rollback or we are only left
with the option to wait for a long time.



Tom Kyte
February 23, 2006 - 10:42 am UTC

that resource busy message means, well, that someone is actually using it, inserting data.

why would you want to truncate their data on them?

truncate or drop

A reader, February 23, 2006 - 11:15 am UTC

I know that no one is logged in so nobody will be inserting
any data in it.



Tom Kyte
February 23, 2006 - 11:24 am UTC

someone is....

else you would not get "resource busy"....


so - look again, there is an open transaction against this table.


select sid, from v$lock where id1 =
(select object_id from dba_objects where owner = TAB_OWNER and object_name = TAB_NAME)
and type = 'TM';


that'll give you the sid's (query v$session to see "who") that have open transactions on this table.

its the killed user

A reader, February 23, 2006 - 11:50 am UTC

Ok, I checked it, its the same USER session whom I
have killed who is having lock in this table, but why
is he still locking it , i have killed it.



Tom Kyte
February 23, 2006 - 6:43 pm UTC

it is not quite dead yet. apparently.

Good stuff

Khalid, February 23, 2006 - 12:56 pm UTC

Response to the last Reader on 2/21 was very informative, especially the HWM thing.

Can we somehow tell Oracle not to apply the roll-forward in the first place, so there is nothing to rollback, that is do not apply the uncommitted stuff from the redo or at least, do not apply anything from the current redo logs?

If shutdown <anykind> is taking forever, and you can't wait, what do you think of doing a 'kill -9' on the pmon, this will surely abort it, just like pulling the plug off, in extremely dire situations of course?


Tom Kyte
February 23, 2006 - 7:09 pm UTC

what roll forward???


If you are talking about recovery (instance recovery) and that would be the only time we would be talking about rolling forward - no, not really. You are thinking too "single user sequentially".

Don't forget that in real life there were 10's, 100's, 1000's or more transactions affecting this same object. These transactions are involved in a complex, interleaved dance with eachother.

If we were a single user, single transaction at a time database - perhaps. But not in real life as it is.

shutdown abort should never hang (well, I know of ONE circumstance where it will error out, but that doesn't count -- anyone guess what it is? will want a demo of course :)

No, do not kill -9 pmon, Yes, do shutdown abort if you want.

No Clue what it would be

Khalid, February 24, 2006 - 11:22 am UTC

I was referring to the roll forward during instance recovery, required after abort. However, I have been reading this same thread a few times, and I have a better understanding.

Now am more curious as to why did you say a shutdown abort would error out? I did a search on otn, I found many candidate errors but my choice is the following:-
ORA-00106 cannot startup/shutdown database when connected to a dispatcher


Tom Kyte
February 24, 2006 - 11:41 am UTC

that isn't the one I was thinking of since you can always connect via a dedicated server to a database to fix that.

I'm thinking of a situation whereby you cannot shutdown abort the database - at all.

Searched google

Khalid, February 24, 2006 - 3:38 pm UTC

One somebody (google hit) had tried to shutdown and he/she couldn't because of some disk full issue, so I theorized as follows:

This is my theory (no POC please), one of the background process is in the middle of writing something crucial to the disk, and the filesystem gets full, so the database just hangs after writing something nasty to the alert log, then I try to shutdown abort and it gives me an error.

But I will be happy to discard my theory if you tell me what it is.


Tom Kyte
February 24, 2006 - 6:37 pm UTC

full file system might do it

moving the dump destination - making that directory inaccessible (which would likely happen with a full file system) does it.

shutdown abort

vivek, March 13, 2006 - 6:34 pm UTC

Tom,

Last week i issued shutdown abort on one instance and then startup the instance. Since then the DB is up and running. Recently i realised that i have not done CLEAN shutdown after startup (last week). Will it cause any problem in future, if i need to bounce the DB again. DB is in no archivelog mode.

Please advise.

Regards

Tom Kyte
March 14, 2006 - 10:03 am UTC

you need not shutdown clean if the database is up and running.

Oracle hangs

Sujith Wimalasooriya, March 15, 2006 - 4:48 pm UTC

Hi Tom,

I apologize if this question is not related to the topic in this thread.

Our database frozen today and I had to reboot the server to get it up and running.
I know that's not the right solution, but after spending 10 minutes with no clue
and the whole company was waiting, I had to.

no new connections were allowed, the active sessions just hangs.
I logged into the server and started a command prompt and couldn't connect.
I started OEM, again on the server, could't connect. No error messages received by
me or any other users. Just frozen.

I looked at the alert.log and nothing bad(ORA- messages) shows up in the last 1 hour or so.

I tried to stop the services before the reboot and it didn't. and then I went for the reboot, which did work.

No other files created in bdump folder in that time frame, other than the alert.log.
There is a file created in udump folder on the same time as the database is restarted, which has a
ORA-28546. I just mentioned it here, but think it was generated when the db was started.
nothing in the cdump. nothing in the bin folder where the sqlplus.exe is. nothing in netwrok\admin folder.
Nothing in sysman\log folder.

From where can I start investigating this case?? we are on windows 2003 server and oracle 9205.

What could I do if this happens again??

Thanks and I appreciate your help.



Tom Kyte
March 16, 2006 - 7:37 am UTC

did you try "connect / as sysdba" on the server, no networking - just a direct connection.



Oracle hangs

Sujith Wimalasooriya, March 15, 2006 - 4:58 pm UTC

Hi Tom,

I missed one information I gathered.
the Server CPU utilization was ~20%.

Thanks,



Oracle hangs

Sujith Wimalasooriya, March 16, 2006 - 9:23 am UTC

Yes, that's what I meant by saying that I tried to logon from the server using the command prompt.

cmd> sqlplus /nolog
sqlplus> connect / as sysdba

This did work in the past and it works even now. But didn't respond back at that time

Thanks,

Oracle hangs

Sujith Wimalasooriya, March 18, 2006 - 8:17 pm UTC

Hi Tom,

Would it be meaningful to enable tracing??
I am not so sure as I couldn't connect from the server, either.

If it happens again, I am going to let it hangs at least for 15 minutes to see if it comes back with a error message.

We have not waited more than 5 minutes before.

Thanks,



Oracle hangs

Sujith Wimalasooriya, April 06, 2006 - 10:26 pm UTC

Hi Tom,

I just deleted the database links we had(They were set up for no real reason) and now it is been three weeks without any problems and thought to share it with others.

I had one ora error and google brought this to my attention,

ORA-28546: connection initialization failed, probable Net8 admin error
Cause: A failure occurred during initialization of a network connection from the Oracle database server to a second process: The connection was completed but a disconnect occurred while trying to perform protocol-specific initialization, usually due to use of different network protocols by opposite sides of the connection. This usually is caused by incorrect Net8 administrative setup for database links or external procedure calls. The most frequent specific causes are:

Database link setup for an Oracle-to-Oracle connection instead connects to a Heterogeneous Services agent or an external procedure agent.
Database link setup for a Heterogeneous Services connection instead connects directly to an Oracle database server.
The extproc_connection_data definition in tnsnames.ora connects to an Oracle instance instead of an external procedure agent.
Connect data for a Heterogeneous Services database link, usually defined in tnsnames.ora, does not specify (HS=).
Connect data for an Oracle-to-Oracle database link, usually defined in tnsnames.ora, specifies (HS=).

Action: Check Net8 administration in the following ways:

When using TNSNAMES.ORA or an Oracle Names server, make sure that the connection from the ORACLE server uses the correct service name or SID.
Check LISTENER.ORA on the connection end point&apos;s host machine to assure that this service name or SID connects to the correct program.
Confirm in TNSNAMES.ORA or the equivalent service definition that service &apos;extproc_connection_data&apos; does NOT contain (HS=), or that the service definition used by a Heterogeneous Services database link DOES contain (HS=).

I didn't understand any of the above, but thought by deleting db links will solve it, and it just did that.

thanks,

Sujith

ORA-01090: shutdown in progress - connection is not permitted

Layla, June 07, 2006 - 8:48 pm UTC

I want to ask about this problem again.
I use SQL PLUS to shutdown the database,and it takes a long time to do it,also it's no responding. So I close this by the task manager. Then the error appeared and all the oracle is not available.
How can I solve this problem.

Tom Kyte
June 08, 2006 - 8:18 am UTC

Well, let's see

you initiate a shutdown (which if you just did "shutdown" will wait until all active sessions log out, but you don't say...)

kill the sqlplus session using task manager...

and then find the database is not available (because it is either A) shutting down or B) shutdown)


Guess I would "start it up"

SYSDBA should be able to connect, use the same credentials you used to initiate the shutdown and do a "startup force"

Is shutdown immediate same as shutdown in 10g

A reader, September 29, 2006 - 2:09 pm UTC

Is shutdown immediate same as shutdown in 10g? I noticed after a shutdown immediate that the database was waiting for active sessions to close. I thought shutdown immediate killed active sessions then did a clean shutdown. Am i just looking at something wrong?

Tom Kyte
September 30, 2006 - 7:54 am UTC

must have been looking at something wrong.


$ sysdba

SQL*Plus: Release 10.2.0.2.0 - Production on Sat Sep 30 07:42:19 2006

Copyright (c) 1982, 2005, Oracle.  All Rights Reserved.


Connected to:
Oracle Database 10g Enterprise Edition Release 10.2.0.2.0 - Production
With the Partitioning, OLAP and Data Mining options

sys%ORA10GR2> select username, status from v$session where username is not null;
USERNAME                       STATUS
------------------------------ --------
OPS$TKYTE                      INACTIVE
SYS                            ACTIVE

sys%ORA10GR2> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
sys%ORA10GR2>


I was the other session (ops$tkyte), and it got killed

 

shutdown immediate by normal user in 9i

Tariq Zia Lakho, November 10, 2006 - 6:20 am UTC

I want to create a normal user called ABC which will work for down the database. What are the rights should I give this user to perform this activity.......................

keeping in mind I am using Oracle9i.
version 9.2.0.1.0


Tom Kyte
November 10, 2006 - 9:02 am UTC

how many places are you going to ask the same thing??????

realize this planet has many timezones and realize I sleep from time to time.

ORA-01089 error over DB Link

Amol, July 16, 2013 - 9:24 am UTC

Dear Tom,

When we try to access a table over DB link from our 9i database we get ORA-01089 error daily.

Then we inform our DBA and he does something and we are able to connect again. Following are error details.

Is it a Oracle bug?

Thanks a lot in advance.

ORA-12012: error on auto execute of job 6 ORA-04052: error occurred when looking up remote object PRD.SYS@PRD.WORLD ORA-00604: error occurred at recursive SQL level 3 ORA-02068: following severe error from ERD ORA-01089: immediate shutdown in progress - no operations are permitted ORA-06512: at "SYS.DBMS_SNAPSHOT", line 803 ORA-06512: at "SYS.DBMS_SNAPSHOT", line 860 ORA-06512: at "SYS.DBMS_IREFRESH", line 683 ORA-06512: at "SYS.DBMS_REFRESH", line 195 ORA-06512: at line 1 ORA-12012: error on auto execute of job 6 ORA-04052: error occurred when looking up remote object PRD.SYS@PRD.WORLD ORA-00604: error occurred at recursive SQL level 3 ORA-02068: following severe error from ERD ORA-01089: immediate shutdown in progress - no operations are permitted ORA-06512: at "SYS.DBMS_SNAPSHOT", line 803 ORA-06512: at "SYS.DBMS_SNAPSHOT", line 860 ORA-06512: at "SYS.DBMS_IREFRESH", line 683 ORA-06512: at "SYS.DBMS_REFRESH", line 195 ORA-06512: at line 1 ORA-12012: error on auto execute of job 6 ORA-04052: error occurred when looking up remote object PRD.SYS@PRD.WORLD ORA-00604: error occurred at recursive SQL level 3 ORA-02068: following severe error from ERD ORA-01089: immediate shutdown in progress - no operations are permitted ORA-06512: at "SYS.DBMS_SNAPSHOT", line 803 ORA-06512: at "SYS.DBMS_SNAPSHOT", line 860 ORA-06512: at "SYS.DBMS_IREFRESH", line 683 ORA-06512: at "SYS.DBMS_REFRESH", line 195 ORA-06512: at line 1 ORA-12012: error on auto execute of job 6 ORA-04052: error occurred when looking up remote object PRD.SYS@PRD.WORLD ORA-00604: error occurred at recursive SQL level 3 ORA-02068: following severe error from ERD ORA-01089: immediate shutdown in progress - no operations are permitted ORA-06512: at "SYS.DBMS_SNAPSHOT", line 803 ORA-06512: at "SYS.DBMS_SNAPSHOT", line 860 ORA-06512: at "SYS.DBMS_IREFRESH", line 683 ORA-06512: at "SYS.DBMS_REFRESH", line 195 ORA-06512: at line 1
Tom Kyte
July 16, 2013 - 4:58 pm UTC

your dba probably is starting the database up again. the remote database is being shutdown.