Lsnrtcl

[harpal]

Hi,

I am a unix person but have a oracle tns listner query, we have a unix server running oracle and the tns listner service we run a backup every night which bribgs down oracle backups the system and then starts the oracle D/B but last night according a successful backup log our oracle D/B was shutdown successfully and restarted successfully but TNS listener failed to start is there any logs which I can check within oracle that will let me ascertain why this was the case I have restarted TNS listener service already.

Thanks

Harpal

 

[Dagon]

Go into lsnrctl and do:

show log_directory
show log_file

There may also be some error messages in your backup log file.

 

[harpal]

Dagon,

I have already checked the log_directory and log_file and had a look at our backup script from last night all fine there jus strange that listener went down.

Thanx

harpal

 

[DBAwhosaysNIE]

hey,

just out of curisoty why are you taking cold backups instead of hot?

there's really no need to ever take down an instance on unix unless you're making init.ora changes (& even some of those can be done hot with alter system) or applying a patch.

 

[harpal]

DBAwhosaysNIE,

Our backup's have always been like this before my time wi the company I do not have much knowledge in oracle, but I would have thought to do hot backups you would need to have rman runnin is this correct or can we do hot backups without rman which is currently not running on our server.

PS no 1 has still not be able give me a solution to my lisenter problem

Thanx

Harpal

 

[DBAwhosaysNIE]

hey,

re: listener problem - if $ORACLE_HOME/network/log/listener.log doesn't contain anything helpful I'd try the system logs (/var/$nix/messages) to see if something happened at the OS level. if that doesn't help just be glad it came back up (I had to reinstall Oracle after a similar problem Monday; see

http://www.tek-tips.com/gviewthread.cfm/lev2/4/lev3/27/pid/186/qid/246923).

re: hot backupe - you most certainly DON'T need rman to do a host backup (I personally don't use ANY Oracle tools like RMAN, OEM, etc). when I used to interview Jr DBAs my rule was if they didn't know how to do it w/SQL*Plus, bash or perl they probably weren't qualified to be doing it. if you'd like I could reply with the scripts (1 bash w/a SQL*Plus subscript) I use.

 

[Turkbear]

Another question:
Are you stopping the listener before the backup? If so, why?

The listener should be running before any instance is started so that the instance can register itself..

If the listener stopped on its own, then the posts about the listener.log are relevant.

 

[spenglerr]

Hi,

what database version of Oracle is used?

Before 8.1.7 you had first to shut down the listener and then the database, and first start the database then the listener.

Since 8.1.7 you will first have to shut the database and then the listener, and then first start the listener and then the database.

Actually you will get no message if you did not configure your listener to trace. So in your $ORACLE_HOME/network/admin/listener.ora-file you will have to include
TRACE_LEVEL_listener_name=0 for doing nothing (Default)
4 for tracing USERS
6 for tracing ADMIN
16 for tracing SUPPORT
will give you details in the $ORACLE_HOME/network/trace/listener.trc-File

If there are severe problems you will find them in $ORACLE_BASE/admin/<db-name>/bdump/<db-name>Alert.log
or $ORACLE_BASE/admin/<db-name>/udump/ORAxxxxx.trc.

Most often problems with TNS arise from DNS-problems. You should have your DNS-entries properly defined for this purpose (especially) at least in the /etc/hosts.

 

[harpal]

we here are runnin ora8i

 

[spenglerr]

Hi,

Ora8i is split into:

8i Release 1 = 8.1.5
8i Release 2 = 8.1.6
8i Release 3 = 8.1.7

when connecting via SQL*Plus the database shows you the version and Patch-Level:

highest:
8.1.6.3 patch 8.1.6.3.8
8.1.7.2 patch 8.1.7.2.5

 

[evatype2]

Assumption: YOU ARE IN A UNIX ENVIRONMENT
Contrary to the information on when you start and stop the listener based on what version of Oracle 8i you are running, you should understand that the listener is a separate process from the database instance. It exists to communicate between the client and the database. I can start and stop a database instance in a UNIX environment without a listener. Some companies like taking the listener down when the database is down because it changes the message a user will get when they try to connect (it is preferable to some to have a request floating around for awhile rather than letting the client listener handle it right away). You can do cold backups with the listener up. The shutdown procedure in your environment prabably has a step to shutdown the listener. Sometimes the database never shuts down correctly. Then the restart fails because the database was still up and then the listener startup step is skipped (depends on how the script is written).

Reboot procedures do need to account for lsnrctl stop and start (again because it is a separate process). Also, the Intelligent Agent needs the listener up and running to start.