[IT Information Online] Kdump - A Powerful Tool for OS Fault Delimitation and Locating

Before using kdump

If the OS breaks down, users can only analyze the cause based on the BMC log and the OS's message log. The useful information is limited.

OS's message log

After using kdump

Users can analyze the cause of OS breakdowns based on the BMC log, OS's message log, and vmcore information collected by Kdump.

kdump generates a vmcore log file at the time the OS restarts after an exception has occurred. As shown in the following example, a kernel bug is recorded in the vmcore log file. This bug has been published on the official Redhat website and the related kernel version is kernel-3.10.0-123.el7.x86_64.

What is kdump?

kdump is a reliable kernel crash dumping mechanism. It was developed by Linux for detecting, collecting, and analyzing kernel crashes. It can be used to find the root cause of crashes and solve critical bugs. The kdump mechanism involves two types of kernel:

- Standard kernel: kernel used for running services

- Capture kernel: kernel specifically used for collecting crash dumps

The standard kernel preserves a small section of memory for kexec to boot the capture kernel. kdump then captures the crash dump image from the context of the freshly booted capture kernel, rather than the crashed standard kernel.

Configuration steps:

1. Configure kdump parameters.

2. Restart the kdump service.

3. Restart the OS for the configuration to take effect.

4. Check whether the configuration is successful.

Notes:

kdump does not affect system performance, but its usage space needs to be configured based on the memory size. For details, see the HUAWEI Server OS Installation Guide.

Advantages:

1. Quick deployment: The deployment is simple and can be finished in about 10 minutes.

2. Quick fault locating: The fault score is narrowed down to help users to identify the fault cause more quickly.

Recommendations:

It is recommended that kdump be configured after OS installation.

For V5 servers, see the installation example for your OS version.

For V2 and V3 servers, see the "Configuring Memory Dump Tools" section in the HUAWEI Server OS Installation Guide.