Skip to content

Test Kdump FADump Configuration

Jump to bottom
Sourabh Jain edited this page Oct 30, 2025 · 2 revisions

Overview

Kdump and FADump are two kernel crash dumping mechanism plays a critical role in First Failure Data Capture (FFDC). The primary job of both Kdump and FADump is to collect the memory dump in the event of a kernel/system crash. The ServiceReport has dedicated plugins that verify the configurations of Kdaump and FADUmp to make sure that the memory dump should be collected whenever there is a kernel crash. Once the configuration of either Kdump or FADump is verified and all the configuration is declared as PASS, it is better to test the workability of the tool before any workload begins its execution. In short, we are advising to take a trial run of these tools. Since testing of Kdump and FADump requires a forceful kernel crash, hence make sure no critical workload is running while initiating the trial run.

Kadump/FADump trail

ServiceReport has a dedicated option -d (--dump) that helps users to trigger the kernel crash if Kdump or FADump is configured correctly. Let's see an example.

$ servicereport -v -d

Based on which kernel crash dumping mechanism is configured in the system, the above command first checks all the configuration of either Kdump or FADumd. If all the configurations are in place the above command triggers the kernel crash. After the kernel crash, the Kdump/FADump tool collects the memory dump (also know as vmcore) and stores it either on the local or remote machine. In all the major Linux distros the default dump-target (place to store memory dump) for Kdump/FADump is /var/crash directory in the local machine. Once the dump is stored at dump-target the system reboots automatically. Now refer the GIF given below to see how to verify whether the memory dump is collected and stored properly or not.

Note: since the dump-target may contain memory dump of old kernel crash hence carefully see the timestamp of every memory dump and then decide whether the recent kernel crash was able to generate the memory dump or not.

Diagram

If the system failed to generate the memory dump (vmcore) file then report this issue on our Github page.

https://github.com/linux-ras/ServiceReport/issues

Please add the following information in the description

  • System memory (free -h)
  • cat /proc/cmdline
  • cat /sys/kernel/kexec_* [if kdump was configured]
  • cat /sys/kernel/fadump_* [if FADump was configured]
Clone this wiki locally