Oracle - Set up automate oracle database startup/shutdown on RHEL7
If we are using chkconfig / init.d for the automate Oracle database shutdown, it will be failed in RHEL7
check the $ORACLE_HOME/shutdown.log, we can find that there is an error for database shutdown by service.
SQL*Plus: Release 12.1.0.2.0 Production on Wed Feb 6 10:29:15 2019 Copyright (c) 1982, 2014, Oracle. All rights reserved. SQL> Connected to an idle instance. SQL> ERROR: ORA-01034: ORACLE not available ORA-27101: shared memory realm does not exist Linux-x86_64 Error: 2: No such file or directory Additional information: 3651 Additional information: -1916159033 Process ID: 0 Session ID: 0 Serial number: 0 SQL> Disconnected Database instance "PPRD3" shut down.
Also check the Alert log, we can see there was no action of database shutdown.
Wed Feb 06 04:15:03 2019 Thread 1 cannot allocate new log, sequence 4438 Private strand flush not complete Current log# 8 seq# 4437 mem# 0: /t03/app/oracle/oradata/PPRD3/redo_AIR_8a.rdo Current log# 8 seq# 4437 mem# 1: /t03/app/oracle/oradata/PPRD3/redo_AIR_8b.rdo Wed Feb 06 04:15:06 2019 Thread 1 advanced to log sequence 4438 (LGWR switch) Current log# 9 seq# 4438 mem# 0: /t03/app/oracle/oradata/PPRD3/redo_AIR_9a.rdo Current log# 9 seq# 4438 mem# 1: /t03/app/oracle/oradata/PPRD3/redo_AIR_9b.rdo Wed Feb 06 10:32:59 2019 Starting ORACLE instance (normal) (OS id: 2266) Wed Feb 06 10:32:59 2019 CLI notifier numLatches:29 maxDescs:2393 Wed Feb 06 10:32:59 2019 ********************************************************************** Wed Feb 06 10:32:59 2019 Dump of system resources acquired for SHARED GLOBAL AREA (SGA) Wed Feb 06 10:32:59 2019 Per process system memlock (soft) limit = 35G Wed Feb 06 10:32:59 2019 Expected per process system memlock (soft) limit to lock SHARED GLOBAL AREA (SGA) into memory: 35G
That means the database is not shutdown gracefully during the machine reboot.
We were using chkconfig to configure service, and in RHEL7, it is converted to systemd internally.
[root@zoro ~]# cat /run/systemd/generator.late/dbora.service # Automatically generated by systemd-sysv-generator [Unit] Documentation=man:systemd-sysv-generator(8) SourcePath=/etc/rc.d/init.d/dbora Description=LSB: Starts the databases on system reboot Before=runlevel3.target Before=runlevel4.target Before=runlevel5.target Before=shutdown.target Before=gcstartup.service Before=lockgcstartup.service After=network-online.target Conflicts=shutdown.target [Service] Type=forking Restart=no TimeoutSec=5min IgnoreSIGPIPE=no KillMode=process GuessMainPID=no RemainAfterExit=yes ExecStart=/etc/rc.d/init.d/dbora start ExecStop=/etc/rc.d/init.d/dbora stop
And in /etc/rc.d/init.d/dbora. it is use “su - oracle -c” to shutdown database.
su - $ORA_OWNER -c "$ORA_HOME/bin/dbshut"
Follow this article:
Automatic Stop of Database (dbshut) not working in OL 7 with systemd (Doc ID 2229679.1)
This is because in RHEL7, init.d is no longer used for the service startup, and it actually be replaced by systemctl. The service which we use chkconfig to create will convert to systemd internally. And dbora script uses su to oracle owner and systemd is not able to keep track of that process.
We can follow
Automate Startup/Shutdown Of Oracle Database 12.2.0.1 On RHEL 7 (Doc ID 2417429.1)
to set up
First, remove chkconfig
# chkconfig --del dbora # systemctl stop dbora.service # systemctl disable dbora.service # systemctl daemon-reload # systemctl reset-failed # systemctl -l|grep dbora <== no entry here
Next, create systemd service
# mkdir /l01/app/oracle/scripts # touch /l01/app/oracle/scripts/start_all.sh #!/bin/bash export ORACLE_SID=12CR1 export ORAENV_ASK=NO . oraenv export ORAENV_ASK=YES lsnrctl start TEST_LISTENER lsnrctl start EXTPROC_LISTENER dbstart # touch /l01/app/oracle/scripts/stop_all.sh #!/bin/bash export ORACLE_SID=12CR1 export ORAENV_ASK=NO . oraenv export ORAENV_ASK=YES lsnrctl stop TEST_LISTENER lsnrctl stop EXTPROC_LISTENER dbshut # chown -R oracle.dba /l01/app/oracle/scripts # chmod u+x /l01/app/oracle/scripts/*.sh # use oracle for testing $ /l01/app/oracle/scripts/start_all.sh $ /l01/app/oracle/scripts/stop_all.sh # touch /lib/systemd/system/dbora.service [Unit] Description=The Oracle Database Service After=syslog.target network.target [Service] # systemd ignores PAM limits, so set any necessary limits in the service. # Not really a bug, but a feature. # https://bugzilla.redhat.com/show_bug.cgi?id=754285 LimitMEMLOCK=infinity LimitNOFILE=65535 LimitNPROC=16384 LimitCORE=0 #Type=simple # idle: similar to simple, the actual execution of the service binary is delayed # until all jobs are finished, which avoids mixing the status output with shell output of services. RemainAfterExit=yes User=oracle Group=dba ExecStart=/l01/app/oracle/scripts/start_all.sh ExecStop=/l01/app/oracle/scripts/stop_all.sh [Install] WantedBy=multi-user.target
before testing, record the setting of limit of SMON process
$ cat /proc/110456/limits Limit Soft Limit Hard Limit Units Max cpu time unlimited unlimited seconds Max file size unlimited unlimited bytes Max data size unlimited unlimited bytes Max stack size 33554432 unlimited bytes Max core file size 0 0 bytes Max resident set unlimited unlimited bytes Max processes 16384 16384 processes Max open files 65536 65536 files Max locked memory 37316722688 37316722688 bytes Max address space unlimited unlimited bytes Max file locks unlimited unlimited locks Max pending signals 2015514 2015514 signals Max msgqueue size 819200 819200 bytes Max nice priority 0 0 Max realtime priority 0 0 Max realtime timeout unlimited unlimited us
# systemctl daemon-reload # systemctl start dbora.service # systemctl enable dbora.service
after testing
- check $ORACLE_HOME/startup.log
- check $ORACLE_HOME/shutdown.log
- check alert log
- compare the ulimit setting.
