DB doesn't go down. [message #322166] |
Thu, 22 May 2008 12:17 |
me_maddy
Messages: 31 Registered: April 2008
|
Member |
|
|
Hello Everybody,
We have issued shutdown immediate towards one of our database.
however it is almost 7 hrs but DB doesn't go down.
below is the Alert log contents.
EMN0 started with pid=13, OS id=9029
Thu May 22 18:10:27 2008
Shutting down instance: further logons disabled
Thu May 22 18:10:28 2008
Stopping background process QMNC
Thu May 22 18:10:28 2008
Stopping background process CJQ0
Thu May 22 18:10:30 2008
Stopping background process MMNL
Thu May 22 18:10:31 2008
Stopping background process MMON
Thu May 22 18:10:32 2008
Shutting down instance (immediate)
License high water mark = 9
Thu May 22 18:10:32 2008
Stopping Job queue slave processes
Thu May 22 18:10:35 2008
Waiting for Job queue slaves to complete
----Thu May 22 18:30:32 2008
KTSMG_UPDATE_MQL(): MMNL absent for 1262 secs; Foregrounds taking over
Listener is already stopped, but There was still one session connected to the DB in addition to SYS user, so we killed that session and oracle displayed the message session marked for kill.
Please suggest what could be the bottleneck.
Thanks
Maddy.
|
|
|
|
Re: DB doesn't go down. [message #322173 is a reply to message #322166] |
Thu, 22 May 2008 12:38 |
me_maddy
Messages: 31 Registered: April 2008
|
Member |
|
|
Quote: | identify the OS process where LOCAL=NO & do kill -9 <pid>
|
I didn't understand the term LOCAL=NO
Please elaborate a bit.
Thanks for your time.
Regards,
Maddy.
|
|
|
|
|
|
Re: DB doesn't go down. [message #322208 is a reply to message #322166] |
Thu, 22 May 2008 15:12 |
me_maddy
Messages: 31 Registered: April 2008
|
Member |
|
|
Thanks Anacedent
I got something now while using the below command.
$ ps -ef |grep LOCAL=YES
ora10 6368 1 0 17:55:13 ? 0:01 oracletnd (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))
Is it safe to kill the pid [6368] now, if yes what would be the impact.
Regards,
Maddy.
|
|
|
|
Re: DB doesn't go down. [message #322212 is a reply to message #322202] |
Thu, 22 May 2008 15:33 |
joy_division
Messages: 4963 Registered: February 2005 Location: East Coast USA
|
Senior Member |
|
|
me_maddy wrote on Thu, 22 May 2008 15:56 | Hi
I tried but didn't get anything
$ ps -ef |grep LOCAL=NO
ora10 28946 8867 0 01:25:01 pts/tg 0:00 grep LOCAL=NO
|
Did you do this on the client, app server or DB server? If you are doing it on your client, this would be the expected result. If you are using an application server, you would need to run that command on the app server.
|
|
|
|
Re: DB doesn't go down. [message #322220 is a reply to message #322216] |
Thu, 22 May 2008 15:55 |
joy_division
Messages: 4963 Registered: February 2005 Location: East Coast USA
|
Senior Member |
|
|
But where do the apps run? Or more precisely, what was connected to the database that you killed?
And you have to clarify what you mean by connected to the DB by telnet. Telnet is a tool to connect to servers, not databases.
[Updated on: Thu, 22 May 2008 15:57] Report message to a moderator
|
|
|
Re: DB doesn't go down. [message #322284 is a reply to message #322216] |
Fri, 23 May 2008 00:23 |
abs_amit
Messages: 59 Registered: January 2006 Location: delhi
|
Member |
|
|
Try this
stop the listener first and then do shutdown abort. After that you can do startup and shutdown immediate for a clean shutdown.
it will rollback all uncompleted transactions.
1. lsnrctl stop <listener_name>
2. sqlplus / as sysdba
SQL> shutdown abort
SQL> startup
SQL> shutdown immediate
Regards
Amit Yadav
|
|
|