How to Completely Disable ‘Kdump’ (Kernel Crash Dumps) in Ubuntu Server

In Ubuntu Server, kdump is a specialized kernel crash dumping mechanism. When a critical system failure (kernel panic) occurs, kdump boots a secondary, lightweight “capture kernel” whose sole purpose is to dump the contents of the system’s memory (RAM) to disk before the machine reboots. This memory dump (vmcore) is invaluable for system administrators and kernel developers to perform post-mortem analysis and determine exactly what caused the crash.

However, running kdump requires permanently reserving a chunk of your system’s RAM—often between 128MB and 512MB—exclusively for the capture kernel. On small Virtual Private Servers (VPS), embedded systems, or cost-optimized cloud instances with limited memory, losing half a gigabyte of RAM to a feature you may never use is highly inefficient. If you do not have the expertise to analyze a kernel dump, or if your server automatically provisions a fresh instance upon failure, you can completely disable kdump to reclaim that memory.

Why Disable Kdump in Ubuntu Server?

Disabling the kernel crash dumping mechanism offers distinct advantages in specific environments:

• Reclaim System RAM: The primary reason to disable kdump is to free up the memory reserved for the capture kernel, allowing your active applications (like web servers or databases) to utilize it.
• Save Disk Space: Kernel memory dumps can be massive—often equivalent to the total amount of RAM installed in the server. Disabling kdump prevents unexpected disk exhaustion during a panic loop.
• Faster Reboot Times: In high-availability clusters, bypassing the memory dump process ensures the server reboots and rejoins the cluster as quickly as possible following a crash.

How to Turn Off ‘Kdump’ in Ubuntu Server

To completely disable kdump, you must modify the GRUB bootloader configuration to remove the memory reservation, and then disable the background service.

1. Connect to your Ubuntu server via SSH or open the terminal.
2. Open the default GRUB configuration file in a text editor with root privileges:
   sudo nano /etc/default/grub
3. Locate the line beginning with GRUB_CMDLINE_LINUX_DEFAULT=.
4. Within the quotes on that line, find the parameter crashkernel=auto (or a specific value like crashkernel=512M).
5. Carefully delete the crashkernel=... parameter. Ensure the rest of the line (like quiet splash) remains intact.
6. Save the file (press Ctrl + O, then Enter) and exit nano (press Ctrl + X).
7. Update the GRUB bootloader to apply the changes to the kernel parameters:
   sudo update-grub
8. Disable the kdump-tools service from starting automatically on boot:
   sudo systemctl disable kdump-tools
9. Reboot your server to boot without the memory reservation:
   sudo reboot

Once the server comes back online, the memory previously reserved for the capture kernel will be fully available to the operating system, and no dumps will be generated upon a kernel panic.