Kernel crashes explained

Author: | Posted in other No comments

Kernel failures, crashes, panics and oopses? Do these sound familiar to you? If not you may want to read dedoimedo’s Linux Kernel Crash Book. It is a step-by-step guide to introduce you to kernel crashes and crash analysis, written with simplicity in mind, accessible even to Linux beginners.

You’ll learn about two kernel crash tools:

LKCD (Linux Kernel Crash Dump)

  • limitations: unable to save memory dumps to local md RAID devices, netdump server must be on the same subnet

Kdump

  • advantages: ia64, ppc64 arch support, sends dumps to a wide range of devices and network shares (NFS, CIFS, FTP, SSH)

The next two chapters describe the crash data collection process and the actual data analysis that is performed to investigate the cause of the crash:

Crash Collection

  • crash setup and running the crash utility

Crash Analysis

  • analyzing the crash reports (backtrace)

You’ll also learn how to:

  • trigger a kernel dump by forcing a kernel panic
  • [code lang=”bash”]echo 1 > /proc/sys/kernel/sysrq ; echo c > /proc/sysrq-trigger[/code]

  • send memory dumps to a remote machine (netdump server)
  • compile a kernel module and use it to create a kernel panic.

Enjoy your reading and don’t forget that submitting kernel crash reports is very important for the developers, as the author points out:

When your kernel crashes, you may want to take the initiative and submit the report to the vendor, so that they may examine it and possibly fix a bug. This is a very important thing. You will not only be helping yourself but possibly everyone using Linux anywhere. What more, kernel crashes are valuable. If there’s a bug somewhere, the developers will find it and fix it.

You can get the book from dedoimedo.com.

Add Your Comment