From Fedora Project Wiki

No edit summary
No edit summary
Line 41: Line 41:


== How To Test ==
== How To Test ==
Enable the following config lines in /etc/kdump.conf
* Dump to local file system
ext3 LABEL=/boot
ext3 /dev/sda1
ext4 UUID=03138356-5e61-4ab3-b58e-27507ac41937
* Dump to remote machines
net my.server.com:/path
net user@remote.server.com
* Dump to raw block device
raw /dev/sda5
* Other configs
core_collector makedumpfile -c --message-level 1
core_collector cp
kdump_post /var/crash/scripts/kdump-post.sh
extra_bins /bin/cp
disk_timeout 30
extra_modules gfs2
options modulename options
default shell
Trigger a crash and see if the vmcore is dumped to the right place.
<!-- This does not need to be a full-fledged document.  Describe the dimensions of tests that this feature is expected to pass when it is done.  If it needs to be tested with different hardware or software configurations, indicate them.  The more specific you can be, the better the community testing can be.  
<!-- This does not need to be a full-fledged document.  Describe the dimensions of tests that this feature is expected to pass when it is done.  If it needs to be tested with different hardware or software configurations, indicate them.  The more specific you can be, the better the community testing can be.  



Revision as of 11:13, 11 July 2011


New mkdumprd for kdump

Summary

This is a replacement for mkdumprd of kexec-tools, to switch to dracut to generate the initrd for kdump kernel.

Owner

  • Email: <amwang at redhat dot com>

Current status

  • Targeted release: Fedora 16
  • Last updated: 2011-07-11
  • Percentage of completion: 10%

Working:

  • Dump to local filesystem

TODO:

  • Read /etc/kdump.conf to generate initrd
  • Dump to remote host via ssh
  • Dump to remote host via nfs
  • Dump to raw block device
  • Implement full functionality of /etc/kdump.conf
  • Dump to multiple targets


Detailed Description

Currently in kexec-tools we use our own /sbin/mkdumprd to generate an initrd for kdump kernel. However, due to the natural complexity of making a new initrd, we have lots of burdens to maintain the mkdumprd code, we decide to switch to using dracut to generate kdump initrd. And hopefully, we can provide a distro-independent mkdumprd to upstream, as dracut itself is cross-distribution.

Benefit to Fedora

The main source of kexec-tools bugs is in mkdumprd, with this change, hopefully we will reduce the number of bugs of kexec-tools.

Scope

All functionality of /sbin/mkdumprd has to be implemented and tested in this new mkdumprd.

How To Test

Enable the following config lines in /etc/kdump.conf

  • Dump to local file system

ext3 LABEL=/boot ext3 /dev/sda1 ext4 UUID=03138356-5e61-4ab3-b58e-27507ac41937

  • Dump to remote machines

net my.server.com:/path net user@remote.server.com

  • Dump to raw block device

raw /dev/sda5

  • Other configs

core_collector makedumpfile -c --message-level 1 core_collector cp kdump_post /var/crash/scripts/kdump-post.sh extra_bins /bin/cp disk_timeout 30 extra_modules gfs2 options modulename options default shell

Trigger a crash and see if the vmcore is dumped to the right place.


User Experience

Users should not notice the replacement, because we plan to provide full backward compatibility.

Dependencies

Dracut

Contingency Plan

Fallback to old /sbin/mkdumprd.

Documentation

  • Not applicable, because we don't need to make documentations for this transparent replacement.

Release Notes

Comments and Discussion

  • Not applicable, because the main discussion is in a Red Hat internal mailing list.