When your Java application suddenly fails with error='Cannot allocate memory' (errno=12), it's typically because the JVM hits a hard memory limit. From the /proc/meminfo output, we can see the server has:

MemTotal: 1027040 kB (~1GB RAM)
SwapTotal: 0 kB (no swap space)

The "it worked yesterday" phenomenon occurs because Linux memory management is dynamic. Your system might have had:

• Other processes releasing memory overnight
• Kernel caches being flushed
• Memory fragmentation issues accumulating

1. Create swap space (temporary solution):

sudo fallocate -l 2G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

2. Reduce JVM heap size:

java -Xms256m -Xmx512m -jar your_application.jar

Configuration for limited-memory servers:

# In application.conf or equivalent:
jvm-memory {
  initial = 256m
  max = 768m
  metaspace = 128m
  stack = 512k
}

Docker memory limits example:

docker run -d \
  -m 1g \
  --memory-swap 1.5g \
  -e JAVA_OPTS="-XX:+UseContainerSupport -XX:MaxRAMPercentage=75.0" \
  your-java-image

Create a heap dump when memory gets critical:

java -XX:+HeapDumpOnOutOfMemoryError \
     -XX:HeapDumpPath=/tmp/heapdump.hprof \
     -XX:+ExitOnOutOfMemoryError \
     -jar your_app.jar

Analyze with jhat or VisualVM:

jhat /tmp/heapdump.hprof
# Then open http://localhost:7000

For JVM internal memory usage:

java -XX:NativeMemoryTracking=detail \
     -XX:+UnlockDiagnosticVMOptions \
     -XX:+PrintNMTStatistics \
     -jar app.jar

Then check allocations:

jcmd <pid> VM.native_memory summary

• Set conservative JVM heap sizes (-Xmx)
• Enable swap space (minimum 1.5x RAM)
• Monitor memory usage with tools like Prometheus
• Consider switching to a more memory-efficient JVM like Eclipse OpenJ9
• Implement proper OOM handling in your application

---

When a Java application that previously ran smoothly suddenly fails with memory allocation errors, it's often a sign of underlying system resource issues. The error message:

Java HotSpot(TM) 64-Bit Server VM warning: INFO: os::commit_memory(0x00000000d5550000, 715849728, 0) failed; error='Cannot allocate memory' (errno=12)

indicates the JVM cannot allocate approximately 715MB of memory. This typically happens when either:

• Physical memory is exhausted
• Swap space is disabled or full
• Process limits are too restrictive

First, verify your current memory situation:

free -h
cat /proc/meminfo
ps aux --sort=-%mem | head -10

In the reported case, we see:

• Total memory: 1GB (1027040 kB)
• Free memory: 626MB
• No swap configured (SwapTotal: 0 kB)

For production systems facing this issue:

# Create swap if none exists (4GB example)
sudo fallocate -l 4G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

Add to /etc/fstab for persistence:

/swapfile none swap sw 0 0

Modify your Java startup parameters:

# Reduce heap size (adjust based on available memory)
java -Xms256m -Xmx512m -jar yourapp.jar

# Alternatively, if using Play Framework:
activator -J-Xms256m -J-Xmx512m -Dhttp.port=80 start

Key JVM memory parameters to consider:

• -XX:MaxDirectMemorySize=256m
• -XX:MaxMetaspaceSize=128m
• -Xss256k (reducing thread stack size)

For persistent issues, enable JVM native memory tracking:

java -XX:NativeMemoryTracking=detail -XX:+UnlockDiagnosticVMOptions \
-XX:+PrintNMTStatistics -Xms512m -Xmx1g yourapp.jar

After running, check memory usage:

jcmd <pid> VM.native_memory summary

Implement basic monitoring with a shell script:

#!/bin/bash
while true; do
  date >> memory.log
  free -m >> memory.log
  ps aux --sort=-%mem | head -5 >> memory.log
  sleep 60
done

For production systems, consider tools like:

• Prometheus + Grafana
• New Relic
• Datadog