Increase Linux ulimit for MYLINEHUB (Fix Too Many Open Files)

In production MYLINEHUB deployments, one of the most common hidden failures is not caused by application bugs — but by Linux resource limits.

When thread or file limits are exhausted, Spring Boot cannot create new threads, background schedulers stop working, and telecom or WhatsApp processing silently fails.

Real Production Error Symptoms

Failed to start thread "Unknown thread" - pthread_create failed (EAGAIN)
Failed to start the native thread for java.lang.Thread "crontaskscheduler-16100"
  

This error means the operating system refused to create a new thread because the process or user has reached its ulimit boundary.

Why This Happens in MYLINEHUB

• High background schedulers (campaigns, WhatsApp, AI jobs)
• Large thread pools for telecom processing
• Default Linux limits too low for production workloads

Without increasing limits, the system appears running but frozen.

Step 1 — Identify Running Java Process

ps aux | grep java

This confirms the active Spring Boot PID and runtime state.

Step 2 — Generate Thread Dump

jstack <PID> > thread-dump.txt

Thread dumps help confirm:

• Thread exhaustion
• Blocked schedulers
• Deadlocks

Step 3 — Check Current Linux Limits

ulimit -u   # max user processes
ulimit -n   # open files
ulimit -s   # stack size
  

Low values here directly cause thread-creation failures.

Step 4 — Increase System Limits Permanently

Edit limits.conf

sudo nano /etc/security/limits.conf

* soft nproc unlimited
* hard nproc unlimited
* soft nofile 65535
* hard nofile 65535
  

Enable PAM limits

sudo nano /etc/pam.d/common-session
session required pam_limits.so
  

Increase kernel file capacity

sudo nano /etc/sysctl.conf
fs.file-max = 2097152

sudo sysctl -p
  

Step 5 — Monitor JVM After Fix

Use built-in JVM tools for real-time monitoring:

• jconsole → threads, memory, GC
• jvisualvm → live profiling & dumps
• jstat → GC and thread statistics

jstat -gcutil <pid> 1000
jstat -thread <pid>
  

How the Problem Looks Before vs After Fix

Before

• Thread creation failures
• Schedulers stop silently
• Calls or WhatsApp jobs freeze

After Increasing ulimit

• Threads create normally
• Schedulers remain stable
• System handles production load smoothly

Production Best Practices

• Always raise nofile and nproc for telecom systems
• Monitor JVM threads using JMX or Actuator
• Use systemd restart policies for fail-safe recovery
• Add Prometheus + Grafana for proactive alerts

Final Thought

Telecom and AI platforms fail silently when system limits are ignored. Increasing ulimit is not tuning — it is a mandatory production requirement for MYLINEHUB stability.