ORA-00845 MEMORY_TARGET Not Supported - Automatic Memory Management Guide

ORA-00845: MEMORY_TARGET Not Supported on This System

Error Overview

Error Text: ORA-00845: MEMORY_TARGET not supported on this system

This critical startup error prevents Oracle database startup when Automatic Memory Management (AMM) cannot allocate the required memory. ORA-00845 typically occurs due to insufficient shared memory filesystem space or incompatible system configuration for AMM.

Understanding Automatic Memory Management

AMM Architecture

Automatic Memory Management (AMM)
├── MEMORY_TARGET (Total memory pool)
├── MEMORY_MAX_TARGET (Maximum memory)
├── System Global Area (SGA)
│   ├── Buffer Cache
│   ├── Shared Pool
│   ├── Large Pool
│   └── Java Pool
└── Program Global Area (PGA)
    ├── Sort Area
    ├── Hash Area
    └── Cursor Cache

Memory Allocation Flow

1. MEMORY_TARGET set → Oracle allocates shared memory
2. Shared memory filesystem (/dev/shm) must have sufficient space
3. Dynamic allocation between SGA and PGA components
4. Operating system must support required memory operations

Common Causes

1. Insufficient /dev/shm Space

• /dev/shm filesystem smaller than MEMORY_TARGET
• Other processes consuming shared memory space
• Incorrect /dev/shm mount configuration
• Container environments with limited shared memory

2. Operating System Limitations

• Older Linux kernels lacking AMM support
• Missing hugepages configuration
• Incompatible shared memory implementation
• SELinux restrictions blocking memory allocation

3. Configuration Mismatches

• MEMORY_TARGET exceeds available system memory
• MEMORY_MAX_TARGET set too high
• Conflicting manual memory parameter settings
• Instance already running with different memory configuration

4. Environment-Specific Issues

• Docker containers without proper shared memory setup
• Cloud instances with restricted memory access
• Virtual machines with insufficient memory allocation
• Container orchestration platforms with memory limits

Diagnostic Steps

1. Check Current Memory Configuration

-- Check memory parameters (if database can connect)
SHOW PARAMETER memory;
SHOW PARAMETER sga;
SHOW PARAMETER pga;

-- Check from SPFILE if database won't start
strings $ORACLE_HOME/dbs/spfile$ORACLE_SID.ora | grep -i memory

2. Verify System Memory and /dev/shm

# Check total system memory
free -h
cat /proc/meminfo | grep MemTotal

# Check /dev/shm filesystem
df -h /dev/shm
mount | grep shm

# Check /dev/shm mount options
cat /proc/mounts | grep shm

# Check available space vs Oracle requirements
ls -la /dev/shm/
du -sh /dev/shm/*

3. Analyze Memory Requirements

# Calculate memory requirements from SPFILE
ORACLE_SID=ORCL  # Set your SID
SPFILE=$ORACLE_HOME/dbs/spfile$ORACLE_SID.ora

echo "Current SPFILE memory settings:"
strings $SPFILE | grep -i memory_target
strings $SPFILE | grep -i memory_max_target
strings $SPFILE | grep -i sga_target

# Check for Oracle shared memory segments
ipcs -m | grep oracle
ipcs -m | grep $ORACLE_USER

4. Check Alert Log and Trace Files

# Check alert log for detailed error information
tail -n 50 $ORACLE_BASE/diag/rdbms/$ORACLE_SID/$ORACLE_SID/trace/alert_$ORACLE_SID.log

# Look for patterns like:
# - "Starting ORACLE instance"
# - "MEMORY_TARGET not supported"
# - "Automatic Memory Management is disabled"
# - "ORA-00845"

# Check for related trace files
ls -la $ORACLE_BASE/diag/rdbms/$ORACLE_SID/$ORACLE_SID/trace/ | grep -E "(ora_|ORA_)" | tail -10

Immediate Solutions

Solution 1: Increase /dev/shm Size

Temporary /dev/shm Resize

# Check current /dev/shm size
df -h /dev/shm

# Calculate required size (MEMORY_TARGET + 20% overhead)
# Example: For MEMORY_TARGET=8G, need at least 10G /dev/shm

# Resize /dev/shm temporarily (requires root)
sudo mount -o remount,size=12G /dev/shm

# Verify new size
df -h /dev/shm

# Attempt Oracle startup
sqlplus / as sysdba
STARTUP;

Permanent /dev/shm Configuration

# Backup current fstab
sudo cp /etc/fstab /etc/fstab.backup

# Edit /etc/fstab (requires root)
sudo vi /etc/fstab

# Add or modify the /dev/shm entry:
tmpfs /dev/shm tmpfs defaults,size=12G 0 0

# Alternative with more explicit options:
tmpfs /dev/shm tmpfs defaults,nodev,nosuid,noexec,size=12G 0 0

# Test the configuration
sudo mount -o remount /dev/shm
df -h /dev/shm

# Restart Oracle
sqlplus / as sysdba
STARTUP;

Solution 2: Disable Automatic Memory Management

Switch to Manual Memory Management

-- Connect using PFILE or SPFILE modification
sqlplus / as sysdba

-- Create PFILE from SPFILE for modification
CREATE PFILE='/tmp/init$ORACLE_SID.ora' FROM SPFILE;

-- Exit and edit PFILE
EXIT;

# Edit the PFILE to disable AMM
vi /tmp/init$ORACLE_SID.ora

# Comment out or remove these lines:
# *.memory_target=8G
# *.memory_max_target=8G

# Add manual memory settings instead:
*.sga_target=6G
*.pga_aggregate_target=2G

# Alternative: Set individual SGA components
# *.sga_target=0
# *.shared_pool_size=1G
# *.db_cache_size=4G
# *.large_pool_size=500M
# *.java_pool_size=200M
# *.pga_aggregate_target=2G

-- Start with PFILE
sqlplus / as sysdba
STARTUP PFILE='/tmp/init$ORACLE_SID.ora';

-- Create new SPFILE with corrected parameters
CREATE SPFILE FROM PFILE='/tmp/init$ORACLE_SID.ora';

-- Restart with new SPFILE
SHUTDOWN IMMEDIATE;
STARTUP;

-- Verify memory configuration
SHOW PARAMETER memory;
SHOW PARAMETER sga;
SHOW PARAMETER pga;

Solution 3: Reduce Memory Allocation

Lower MEMORY_TARGET Values

-- If database starts intermittently, reduce memory allocation
sqlplus / as sysdba

-- Reduce memory targets
ALTER SYSTEM SET memory_target = 6G SCOPE = SPFILE;
ALTER SYSTEM SET memory_max_target = 6G SCOPE = SPFILE;

-- Restart database
SHUTDOWN IMMEDIATE;
STARTUP;

Calculate Optimal Memory Settings

#!/bin/bash
# memory_calculator.sh - Calculate optimal Oracle memory settings

# Get system memory in GB
TOTAL_MEM_KB=$(grep MemTotal /proc/meminfo | awk '{print $2}')
TOTAL_MEM_GB=$((TOTAL_MEM_KB / 1024 / 1024))

echo "System Memory Analysis:"
echo "Total System Memory: ${TOTAL_MEM_GB}GB"

# Calculate recommendations
OS_RESERVE_GB=2
OTHER_APPS_GB=1
AVAILABLE_GB=$((TOTAL_MEM_GB - OS_RESERVE_GB - OTHER_APPS_GB))

# Oracle memory recommendations (70% of available)
ORACLE_MEM_GB=$((AVAILABLE_GB * 70 / 100))
SGA_GB=$((ORACLE_MEM_GB * 75 / 100))
PGA_GB=$((ORACLE_MEM_GB * 25 / 100))

echo ""
echo "Recommended Oracle Memory Settings:"
echo "MEMORY_TARGET: ${ORACLE_MEM_GB}G"
echo "Or Manual Settings:"
echo "SGA_TARGET: ${SGA_GB}G"
echo "PGA_AGGREGATE_TARGET: ${PGA_GB}G"
echo ""
echo "Required /dev/shm size: $((ORACLE_MEM_GB + 1))G"

# Check current /dev/shm
SHM_SIZE_KB=$(df /dev/shm | awk 'NR==2 {print $2}')
SHM_SIZE_GB=$((SHM_SIZE_KB / 1024 / 1024))

echo "Current /dev/shm size: ${SHM_SIZE_GB}G"

if [ $SHM_SIZE_GB -lt $((ORACLE_MEM_GB + 1)) ]; then
    echo "WARNING: /dev/shm too small. Need to increase to $((ORACLE_MEM_GB + 1))G"
else
    echo "OK: /dev/shm size is sufficient"
fi

Long-Term Solutions

1. Container Environment Configuration

Docker Configuration

# Dockerfile for Oracle with proper shared memory
FROM oracle/database:19.3.0-ee

# Configure shared memory in container
RUN echo "tmpfs /dev/shm tmpfs defaults,size=12G 0 0" >> /etc/fstab

# Set Oracle memory parameters
ENV ORACLE_MEMORY_TARGET=8G
ENV ORACLE_SGA_TARGET=6G
ENV ORACLE_PGA_TARGET=2G

# Run Oracle container with increased shared memory
docker run -d \
  --name oracle-db \
  --shm-size=12g \
  -e ORACLE_SID=ORCL \
  -e ORACLE_PDB=PDB1 \
  -e ORACLE_PWD=YourPassword \
  -e ORACLE_MEMORY_TARGET=8G \
  -p 1521:1521 \
  -p 5500:5500 \
  oracle/database:19.3.0-ee

# Check container shared memory
docker exec oracle-db df -h /dev/shm

Kubernetes Configuration

kubernetes-oracle.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: oracle-database
spec:
  replicas: 1
  selector:
    matchLabels:
      app: oracle-database
  template:
    metadata:
      labels:
        app: oracle-database
    spec:
      containers:
      - name: oracle
        image: oracle/database:19.3.0-ee
        env:
        - name: ORACLE_SID
          value: "ORCL"
        - name: ORACLE_PWD
          value: "YourPassword"
        - name: ORACLE_MEMORY_TARGET
          value: "8G"
        resources:
          requests:
            memory: "12Gi"
            cpu: "2"
          limits:
            memory: "16Gi"
            cpu: "4"
        volumeMounts:
        - name: dshm
          mountPath: /dev/shm
        - name: oracle-data
          mountPath: /opt/oracle/oradata
      volumes:
      - name: dshm
        emptyDir:
          medium: Memory
          sizeLimit: 12Gi
      - name: oracle-data
        persistentVolumeClaim:
          claimName: oracle-data-pvc

2. System Optimization for AMM

Configure Hugepages for Better Performance

# Calculate hugepages needed
MEMORY_TARGET_GB=8
HUGEPAGE_SIZE_KB=$(grep Hugepagesize /proc/meminfo | awk '{print $2}')
HUGEPAGE_SIZE_MB=$((HUGEPAGE_SIZE_KB / 1024))
HUGEPAGES_NEEDED=$((MEMORY_TARGET_GB * 1024 / HUGEPAGE_SIZE_MB))

echo "Hugepages calculation:"
echo "Memory Target: ${MEMORY_TARGET_GB}GB"
echo "Hugepage Size: ${HUGEPAGE_SIZE_MB}MB"
echo "Hugepages Needed: ${HUGEPAGES_NEEDED}"

# Set hugepages
sudo sysctl -w vm.nr_hugepages=$HUGEPAGES_NEEDED

# Make permanent
echo "vm.nr_hugepages = $HUGEPAGES_NEEDED" | sudo tee -a /etc/sysctl.conf

# Configure Oracle to use hugepages
sqlplus / as sysdba
ALTER SYSTEM SET use_large_pages = TRUE SCOPE = SPFILE;

Memory Monitoring and Alerting

-- Create memory monitoring procedure
CREATE OR REPLACE PROCEDURE check_memory_config AS
  v_memory_target NUMBER;
  v_sga_size NUMBER;
  v_pga_size NUMBER;
  v_shm_available NUMBER;
BEGIN
  -- Get current memory configuration
  SELECT value INTO v_memory_target
  FROM v$parameter WHERE name = 'memory_target';

  SELECT SUM(value) INTO v_sga_size FROM v$sga;

  SELECT value INTO v_pga_size
  FROM v$pgastat WHERE name = 'total PGA allocated';

  -- Check if AMM is enabled
  IF v_memory_target > 0 THEN
    DBMS_OUTPUT.PUT_LINE('AMM enabled with MEMORY_TARGET: ' ||
                        ROUND(v_memory_target/1024/1024/1024, 2) || 'GB');

    -- Check /dev/shm space (would need external script)
    DBMS_OUTPUT.PUT_LINE('Verify /dev/shm has sufficient space');
  ELSE
    DBMS_OUTPUT.PUT_LINE('Manual memory management in use');
  END IF;

  DBMS_OUTPUT.PUT_LINE('Current SGA: ' || ROUND(v_sga_size/1024/1024/1024, 2) || 'GB');
  DBMS_OUTPUT.PUT_LINE('Current PGA: ' || ROUND(v_pga_size/1024/1024/1024, 2) || 'GB');
END;
/

-- Run the check
EXEC check_memory_config;

3. Cloud Environment Best Practices

AWS RDS and EC2 Configuration

# For AWS EC2 instances running Oracle
# Choose memory-optimized instances (r5, r6i, x1e)

# Instance type recommendations:
# r5.2xlarge: 64GB RAM - supports up to 40GB Oracle memory
# r5.4xlarge: 128GB RAM - supports up to 80GB Oracle memory
# r5.8xlarge: 256GB RAM - supports up to 180GB Oracle memory

# Configure /dev/shm in user data script
#!/bin/bash
echo "tmpfs /dev/shm tmpfs defaults,size=50G 0 0" >> /etc/fstab
mount -o remount /dev/shm

# Set optimal kernel parameters
echo "vm.nr_hugepages = 20480" >> /etc/sysctl.conf  # 40GB hugepages
sysctl -p

Azure and GCP Configuration

# Azure VM configuration
# Use memory-optimized series (Dv3, Ev3, Mv2)

# GCP configuration
# Use memory-optimized machine types (n1-highmem, n2-highmem)

# Common cloud configuration script
#!/bin/bash
# cloud_oracle_setup.sh

# Get total memory
TOTAL_MEM_GB=$(free -g | awk 'NR==2{print $2}')
ORACLE_MEM_GB=$((TOTAL_MEM_GB * 60 / 100))  # 60% for Oracle
SHM_SIZE_GB=$((ORACLE_MEM_GB + 5))          # Oracle memory + 5GB buffer

echo "Configuring Oracle for ${TOTAL_MEM_GB}GB system"
echo "Oracle memory: ${ORACLE_MEM_GB}GB"
echo "/dev/shm size: ${SHM_SIZE_GB}GB"

# Configure /dev/shm
cp /etc/fstab /etc/fstab.backup
echo "tmpfs /dev/shm tmpfs defaults,size=${SHM_SIZE_GB}G 0 0" >> /etc/fstab
mount -o remount /dev/shm

# Configure hugepages if memory > 16GB
if [ $ORACLE_MEM_GB -gt 16 ]; then
    HUGEPAGES=$((ORACLE_MEM_GB * 512))  # 2MB hugepages
    echo "vm.nr_hugepages = $HUGEPAGES" >> /etc/sysctl.conf
    sysctl -p
fi

echo "Oracle memory configuration completed"

Prevention Strategies

1. Pre-deployment Validation

#!/bin/bash
# validate_oracle_memory.sh - Pre-deployment memory validation

REQUIRED_MEMORY_GB=${1:-8}
ORACLE_SID=${2:-ORCL}

echo "Validating memory configuration for Oracle $ORACLE_SID"
echo "Required memory: ${REQUIRED_MEMORY_GB}GB"

# Check system memory
TOTAL_MEM_GB=$(free -g | awk 'NR==2{print $2}')
if [ $TOTAL_MEM_GB -lt $((REQUIRED_MEMORY_GB * 2)) ]; then
    echo "ERROR: Insufficient system memory"
    echo "Required: $((REQUIRED_MEMORY_GB * 2))GB, Available: ${TOTAL_MEM_GB}GB"
    exit 1
fi

# Check /dev/shm
SHM_SIZE_GB=$(df /dev/shm | awk 'NR==2 {print int($2/1024/1024)}')
if [ $SHM_SIZE_GB -lt $((REQUIRED_MEMORY_GB + 2)) ]; then
    echo "ERROR: Insufficient /dev/shm space"
    echo "Required: $((REQUIRED_MEMORY_GB + 2))GB, Available: ${SHM_SIZE_GB}GB"
    echo "Fix: sudo mount -o remount,size=$((REQUIRED_MEMORY_GB + 2))G /dev/shm"
    exit 1
fi

# Check for existing Oracle processes
if pgrep -f "ora_.*_$ORACLE_SID" > /dev/null; then
    echo "WARNING: Oracle instance $ORACLE_SID appears to be running"
fi

echo "Memory validation completed successfully"

2. Automated Memory Management Scripts

#!/bin/bash
# oracle_memory_manager.sh - Automated memory management

ORACLE_SID=${1:-ORCL}
ACTION=${2:-check}

case $ACTION in
    "check")
        echo "Checking Oracle memory configuration..."
        df -h /dev/shm
        free -h
        ipcs -m | grep oracle
        ;;

    "increase-shm")
        NEW_SIZE=${3:-16G}
        echo "Increasing /dev/shm to $NEW_SIZE..."
        sudo mount -o remount,size=$NEW_SIZE /dev/shm
        df -h /dev/shm
        ;;

    "fix-amm")
        echo "Attempting to fix AMM configuration..."

        # Check if database is running
        if pgrep -f "ora_smon_$ORACLE_SID" > /dev/null; then
            echo "Database is running, checking memory usage..."
            sqlplus -s / as sysdba << EOF
SET PAGESIZE 0 FEEDBACK OFF
SELECT 'Memory Target: ' || value/1024/1024/1024 || 'GB' FROM v\$parameter WHERE name='memory_target';
SELECT 'SGA Size: ' || SUM(value)/1024/1024/1024 || 'GB' FROM v\$sga;
EXIT;
EOF
        else
            echo "Database is not running, checking /dev/shm..."
            df -h /dev/shm

            # Attempt startup
            sqlplus / as sysdba << EOF
STARTUP;
EXIT;
EOF
        fi
        ;;

    *)
        echo "Usage: $0 <ORACLE_SID> <check|increase-shm|fix-amm> [size]"
        exit 1
        ;;
esac

Related Errors

• ORA-27125 - Unable to create shared memory segment
• ORA-04031 - Unable to allocate shared memory
• ORA-00600 - Internal error (memory corruption)
• ORA-01034 - Oracle not available (startup failure)

Emergency Response

Quick Recovery Steps

1. Check /dev/shm space immediately

   df -h /dev/shm

2. Increase /dev/shm if needed

   sudo mount -o remount,size=16G /dev/shm

3. Try database startup

   sqlplus / as sysdba
   STARTUP;

4. If still failing, disable AMM

   CREATE PFILE='/tmp/init.ora' FROM SPFILE;
   -- Edit PFILE to remove memory_target
   STARTUP PFILE='/tmp/init.ora';

Quick Reference Commands

# Check memory status
df -h /dev/shm
free -h
ipcs -m | grep oracle

# Increase /dev/shm
sudo mount -o remount,size=16G /dev/shm

# Oracle memory commands
sqlplus / as sysdba
SHOW PARAMETER memory;
ALTER SYSTEM SET memory_target = 6G SCOPE = SPFILE;
STARTUP;

# Disable AMM emergency
CREATE PFILE FROM SPFILE;
# Edit PFILE: comment out memory_target, add sga_target/pga_aggregate_target
STARTUP PFILE='/path/to/pfile';

System Configuration Checklist

# Essential /etc/fstab entry
tmpfs /dev/shm tmpfs defaults,size=16G 0 0

# Essential /etc/sysctl.conf entries (for hugepages)
vm.nr_hugepages = 8192
vm.hugetlb_shm_group = 54321  # Oracle group ID

# Verify configuration
mount -o remount /dev/shm
sysctl -p
df -h /dev/shm