Force a service into offline mode
Michael Schmarck
This morning, I noticed the following:
,----[ non functional oracle service ]
| --($:/opt/apps/Teamcenter/tcpre51/software.6/logs)-- svcs -a | grep \*
| offline* 3:43:55 svc:/application/oracle/database:RACE002
|
| --($:/opt/apps/Teamcenter/tcpre51/software.6/logs)-- svcs -xv svc:/application/oracle/database:RACE002
| svc:/application/oracle/database:RACE002 (Oracle database)
| State: offline since 8. Januar 2008 03:43:55 CET
| Reason: Start method is running.
| See: http://sun.com/msg/SMF-8000-C4
| See: man -M /usr/share/man -s 5 ora-smf
| See: /var/svc/log/application-oracle-database:RACE002.log
| Impact: This service is not running.
|
| --($:/opt/apps/Teamcenter/tcpre51/software.6/logs)-- tail /var/svc/log/application-oracle-database:RACE002.log
| database RACE002 is OPEN.
| [ Jan 7 03:41:45 Method "start" exited with status 0 ]
| [ Jan 8 03:42:52 Stopping because service disabled. ]
| [ Jan 8 03:42:52 Executing stop method ("/lib/svc/method/ora-smf stop database RACE002") ]
| Database closed.
| Database dismounted.
| ORACLE instance shut down.
| [ Jan 8 03:43:00 Method "stop" exited with status 0 ]
| [ Jan 8 03:43:54 Enabled. ]
| [ Jan 8 03:43:55 Executing start method ("/lib/svc/method/ora-smf start database RACE002") ]
|
| --($:~)-- LC_ALL=C date
| Tue Jan 8 07:52:52 CET 2008
`----
This means, that the oracle/database service for instance RACE002 did
not come up - the "Start method is running" all the time since about
4am until now.
What can I do, to force the service "svc:/application/oracle/database:RACE002"
into offline mode? I tried running
svcadm disable svc:/application/oracle/database:RACE002
But this did not seem to have any effect, whatsoever.
Thanks a lot,
Michael
victorf...@yahoo.com
If you do not have a DBA, you can log in as oracle and do following
oracle@hostname> . oraenv
ORACLE_SID = [RACE002] ? (if it is not RACE002, then type RACE002)
oracle@hostname> sqlplus /nolog
SQL> conn / as sysdba
Connected.
SQL> shutdown
SQL> startup
SQL> quit
if the startup failed, you need to check alert_RACE002.log which could
be founded by searching $ORACLE_BASE directory ($ORACLE_BASE/admin/
race002/bdump by default).
Victor
Michael Schmarck
Well, thanks, but this question was not intended to be about Oracle :)
I'd like to know, how to make Solaris behave - how to force a service
into offline mode in Solaris.
Thanks a lot though :-)
Michael
Michael Schmarck
Hello again!
This, sort of, happened to me again. On the system, a process required
too much memory and because of that, some processes died - one of these
was tnslsnr from Oracle - the Oracle LISTENER.
,----[ hung oracle listener service ]
| $ svcs svc:/application/oracle/listener:LISTENER
| STATE STIME FMRI
| online* 15:31:08 svc:/application/oracle/listener:LISTENER
|
| $ svcs -xv svc:/application/oracle/listener:LISTENER
| svc:/application/oracle/listener:LISTENER (Oracle listener)
| Status: online seit 2. April 2008 15:31:08 CEST
| Siehe: man -M /usr/share/man -s 5 ora-smf
| Siehe: /var/svc/log/application-oracle-listener:LISTENER.log
| Auswirkung: Keine.
|
| $ tail /var/svc/log/application-oracle-listener:LISTENER.log
| The command completed successfully
|
| LSNRCTL for Solaris: Version 9.2.0.4.0 - Production on 02-APR-2008 15:31:29
|
| Copyright (c) 1991, 2002, Oracle Corporation. All rights reserved.
|
| Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=winds06.win.ch.da.rtr)(PORT=1521)))
| The command completed successfully
| listener LISTENER stop succeeded
| [ Apr 2 15:31:29 Method "stop" exited with status 0 ]
`----
As you can see, the process is in "online*", ie. in becoming
online. But it just hangs there.
I tried to disable the service by doing
"svcadm disable svc:/application/oracle/listener:LISTENER", but this
didn't change anything at all - the process is still "online*".
What can I do, so that the service moves on to "online" or "offline"
or "maintenance"?
I'm using Oracle SMF from http://joostm.nl/, if that matters; the
ora-smf-head.pkg to be exact.
Thanks a lot,
Michael
PS: I was able to restart the listener by doing "lsnrctl start". So
the Oracle problem is solved. I'm now looking for a way to make
Solaris 10 U4 Sparc (with all the latest patches) behave the way I
want and need it to.