# Understanding High I/O Wait

Before we dive into fixes, let’s understand what “I/O wait” means.

• I/O wait is the percentage of time your CPU spends waiting for disk input/output operations (like reading/writing to disk).

• High I/O wait doesn’t mean your CPU is busy — it’s idle, stuck waiting for the disk.

• Common causes:

  • Slow hard drives (HDD vs SSD)

  • Too many processes reading/writing at the same time

  • Database intensive queries

  • Log flooding or backups running

  • Swap usage due to low RAM

  • 
    

---

## Step 1 — Check CPU & I/O Usage

Use top or htop to check I/O wait.

top

Look for %wa under the CPU row. Example:

Cpu(s): 10.0%us, 5.0%sy, 80.0%wa, 0.0%id

• %wa = 80% → CPU is mostly waiting for disk.

Or use htop and add I/O wait column.

---

## Step 2 — Identify Processes Causing High I/O

Use iotop (install if necessary):

sudo apt install iotop -y      # Debian/Ubuntu
sudo yum install iotop -y      # CentOS/RHEL
sudo iotop -ao

Look for:

• Processes using the most DISK READ or DISK WRITE

• Common culprits: mysqld, php-fpm, backup scripts, rsync, logrotate, docker containers

• 
  

---

## Step 3 — Check Disk Health

If your disk is slow, everything waits.

sudo smartctl -a /dev/sda      # check disk health (SMART)
sudo iostat -dx 1               # shows I/O stats per device

Look for:

• r/s (reads per second)

• w/s (writes per second)

• await (average wait per I/O)

• High await → disk is bottleneck

• 
  

---

## Step 4 — Check Swap Usage

If RAM is low, Linux swaps memory to disk, increasing I/O wait.

free -h
swapon --show

High swap usage → server constantly reading/writing disk → CPU waits.

Fix:

• Add RAM

• Reduce memory-hungry processes

• Disable unnecessary services

• 
  

---

## Step 5 — Optimize Logs and Temporary Files

Many logs, especially for web servers, can cause constant disk writes.

Check log size:

du -sh /var/log/*

Truncate or rotate logs:

sudo truncate -s 0 /var/log/syslog
sudo truncate -s 0 /var/log/mysql/error.log
sudo logrotate -f /etc/logrotate.conf

---

## Step 6 — Optimize Database I/O

If MySQL/PostgreSQL is writing too much:

• Add proper indexes to reduce full table scans

• Tune buffer/cache size

• Use SSD instead of HDD if possible

• Reduce frequency of heavy queries or batch jobs

Check MySQL I/O:

SHOW GLOBAL STATUS LIKE 'Innodb_data_%';

---

## Step 7 — Reduce Heavy Disk Usage Scripts

Identify cron jobs or scripts writing lots of data.

ps aux --sort=-%mem | head -n 10

Or check disk write rate with iotop.

Common fixes:

• Batch data processing in smaller chunks

• Move temporary files to faster storage (SSD or tmpfs)

• Avoid multiple simultaneous heavy jobs

• 
  

---

## Step 8 — Use Faster Storage (if needed)

• Upgrade from HDD → SSD

• Move database or logs to separate disk

• Use RAID 10 for high write I/O

• Consider cloud storage with better IOPS (AWS EBS io1/io2, DigitalOcean volumes)

• 
  

---

## Step 9 — Tune Linux I/O Scheduler

Linux uses I/O schedulers like cfq, deadline, noop. For SSDs:

cat /sys/block/sda/queue/scheduler

Recommended for SSD: deadline or noop.

Change temporarily:

echo deadline | sudo tee /sys/block/sda/queue/scheduler

---

## Step 10 — Monitor & Prevent Future High I/O Wait

• Install monitoring tools: Netdata, Glances, Prometheus + Grafana

• Check /var/log/syslog for repetitive errors

• Schedule heavy jobs during low-traffic hours

• Consider caching solutions like Redis or Memcached to reduce disk access

• 
  

---

## Beginner-Friendly Checklist

---

## Conclusion

High I/O wait is a sign that your server is bottlenecked by disk operations. By following this step-by-step guide, even beginners can:

• Identify which processes or services cause high I/O

• Check disk health and usage

• Optimize logs, scripts, and databases

• Upgrade storage or tune I/O scheduler

• Monitor and prevent future issues

With these steps, your Linux server should feel responsive again, and you’ll understand the root cause instead of just rebooting blindly.