Oracle DBA

Linux patching (OS security updates, kernel upgrades, etc.) is a routine activity for system administrators. For Oracle RAC environments, DBAs must ensure pre-patching checks, safe shutdowns, and post-patch relinks are properly executed to avoid service disruption.

---

1. Why Pre-Patching Checks Are Critical

Before handing over the server to the Linux team, DBAs must:

• Capture database and listener state
• Record clusterware resources
• Collect ASM diskgroup details
• Backup OS and crontab configurations

These pre-checks serve as a baseline snapshot. After patching, DBAs can compare the current state to confirm successful recovery.

---

2. Pre-Patching Activities

OS-Level Checks

Run these commands as oracle/grid users to capture environment details:

# Database processes
ps -eaf | grep pmon
ps -eaf | grep tns

# Mounted filesystems
df -kh

# Oracle inventory files
cat /etc/oratab
cat /etc/hosts

# Crontab backup
crontab -l > crontab_backup.txt

# Listener status
lsnrctl status LISTENER

Database-Level Checks

Connect to each database and run:

-- Database role and mode
select name, open_mode, database_role, controlfile_type, switchover_status, db_unique_name 
from gv$database order by 1;

-- SPFILE location
show parameter spfile;

-- Controlfile details
show parameter control;

-- Listener, service, cluster parameters
show parameter listener;
show parameter service;
show parameter cluster;

-- Datafile and tempfile
select name from v$datafile;
select name from v$tempfile;

-- Redo log files
select member from v$logfile;

-- Controlfile locations
select name from v$controlfile;

-- ASM Diskgroup info
select name, state, type, total_mb, free_mb from v$asm_diskgroup;
select name, path, header_status, mode_status from v$asm_disk;

Low-Level Disk Checks

# Device files
cd /dev
ls -larth

# Multipath devices
cd /dev/mapper
ls -larth

Grid Infrastructure Checks

Run as grid user:

# Grid processes
ps -eaf | grep grid

# Clusterware resource status
crsctl stat res -t

# CRS health check
crsctl check crs

# Listener status
lsnrctl status LISTENER

👉 Important: Before patching, disable CRS to prevent automatic startup during reboot.

su - root
$GRID_HOME/bin/crsctl disable crs

This ensures CRS does not start accidentally while Linux team is still patching or rebooting nodes.

---

3. Automated Pre-Patch Script

Here’s a ready-to-use script to gather all pre-patching data:

#!/bin/bash
# prepatch_check.sh
# Run as oracle/grid before handing server to Linux team

LOGFILE=prepatch_$(hostname)_$(date +%Y%m%d%H%M).log

echo "===== Pre-Patching Checks on $(hostname) =====" | tee -a $LOGFILE

echo "== Database Processes ==" | tee -a $LOGFILE
ps -eaf | grep pmon | tee -a $LOGFILE
ps -eaf | grep tns | tee -a $LOGFILE

echo "== Filesystems ==" | tee -a $LOGFILE
df -kh | tee -a $LOGFILE

echo "== /etc/oratab ==" | tee -a $LOGFILE
cat /etc/oratab | tee -a $LOGFILE

echo "== /etc/hosts ==" | tee -a $LOGFILE
cat /etc/hosts | tee -a $LOGFILE

echo "== Crontab ==" | tee -a $LOGFILE
crontab -l | tee crontab_backup_$(whoami).txt | tee -a $LOGFILE

echo "== Listener Status ==" | tee -a $LOGFILE
lsnrctl status LISTENER | tee -a $LOGFILE

echo "== CRS Status ==" | tee -a $LOGFILE
crsctl stat res -t | tee -a $LOGFILE
crsctl check crs | tee -a $LOGFILE

echo "== Device Mapping ==" | tee -a $LOGFILE
ls -larth /dev | tee -a $LOGFILE
ls -larth /dev/mapper | tee -a $LOGFILE

echo "Pre-patch checks completed. Logs saved at $LOGFILE"

Run:

chmod +x prepatch_check.sh
./prepatch_check.sh

---

4. Handover to Linux Team

At this stage:

• DBs and listeners should be stopped
• Crontabs are backed up
• CRS disabled (crsctl disable crs)

Now handover the server to the Linux team for patching and reboot.

---

5. Post-Patching Activities

After Linux patching is complete, perform the following per RAC node:

Step 1: Relink Grid Infrastructure

su - root
$GRID_HOME/crs/install/rootcrs.sh -unlock

su - grid
cd $GRID_HOME/bin
relink all

su - root
$GRID_HOME/crs/install/rootcrs.sh -lock
$GRID_HOME/bin/crsctl enable crs
$GRID_HOME/bin/crsctl start crs

Step 2: Relink Oracle Database

su - oracle
cd $ORACLE_HOME/bin
relink all

Step 3: Startup and Verification

-- Start databases
sqlplus / as sysdba
startup;

-- Verify role/mode again
select name, open_mode, database_role from gv$database;

-- Check listeners
lsnrctl status LISTENER;

Step 4: Restore Crontabs

su - oracle
crontab oracle.cron

su - grid
crontab grid.cron

---

6. Conclusion

By following this pre-check, handover, relink, and verification cycle, DBAs ensure:

• Safe patching coordination with Linux admins
• No surprises with missing processes or CRS failures after reboot
• Grid and Database binaries properly relinked against the new Linux libraries
• Database services fully restored with baseline checks verified

This approach minimizes risk and ensures smooth Linux patching in Oracle RAC environments.