Ansible Role: sys_kdump

An Ansible to install and configure kdump on a host.

Motivation

Kdump and kexec are tools often used to debug and analyze errors and crashes.

Description

This role ensure, that kexec and kdump are properly installed and configured.

Requirements

Used Modules:

Installation

Install from Ansible Galaxy

ansible-galaxy install while_true_do.sys_kdump

Install from Github

git clone https://github.com/while-true-do/ansible-role-sys_kdump.git while_true_do.sys_kdump

Usage

Role Variables

---
# defaults file for while_true_do.sys_kdump

## Package Management
wtd_sys_kdump_package:
  - kexec-tools
# State can be present|latest|absent
wtd_sys_kdump_package_state: "present"

## Service Management
wtd_sys_kdump_service: "kdump"
# State can be started|stopped
wtd_sys_kdump_service_state: "started"
wtd_sys_kdump_service_enabled: true

## Configuration Management
# Variables for /etc/kdump.conf
# Please read templates/kdump.conf.j2
wtd_sys_kdump_conf:
  # Define the method to dump the vmcore
  # (raw|nfs|ssh|<fs type>|path)
  method: "path"
  # address or path for method
  target: "/var/crash"
  # ssh key, in case you want to use method: ssh^
  # sshkey: "/root/.ssh/kdump_id_rsa"
  # parameters for collector
  # collector: "makedumpfile -F -l --message-level 1 -d 31"
  collector: "makedumpfile -l --message-level 1 -d 31"
  # binary|script to run after vmcore dump
  # post: "/var/crash/scripts/kdump-post.sh"
  # binary|script to run before vmcore dump
  # pre: "/var/crash/scripts/kdump-pre.sh"
  # binary|script to be included in the kdump initrd
  # extra_bins: "/usr/bin/lftp"
  # extra kernel modules included in kdump initrd
  # extra_modules: "gfs2"
  # action to perform, in case kdump fails
  # (reboot|halt|poweroff|shell)
  default: "reboot"
  # specify, if initrd must be rebuild
  # (0|1)
  # force_rebuild: "1"
  # specify, if initrd must never be rebuild
  # (0|1)
  # force_no_rebuild: "1"
  # force unresettable blockdevice as target
  # (0|1)
  # override_resettable: "1"
  # extra dracut options to be inclued in initrd
  # dracut_args: '--omit-drivers "cfg80211 snd" --add-drivers "ext2 ext3"'
  # arguments for fence_kdump_send
  # fence_kdump_args: "-p 7410 -f auto -c 0 -i 10"
  # list all cluster nodes, except localhost
  # fence_kdump_nodes: "node1 node2"

## Host Management
wtd_sys_kdump_reboot_enabled: false
wtd_sys_kdump_reboot_msg: "System is going down to apply kdump configuration."
wtd_sys_kdump_reboot_timeout: "3600"

Example Playbook

Running Ansible Roles can be done in a playbook.

Simple

---
- hosts: all
  roles:
    - role: while_true_do.sys_kdump

Advanced

Use an nfs address as target.

- hosts: all
  roles:
    - role: while_true_do.sys_kdump
      wtd_sys_kdump_conf:
        method: "nfs"
        target: "192.168.10.1:/path/to/dir"
        collector: "makedumpfile -l --message-level 1 -d 31"
        default: "reboot"

Known Issues

RedHat Testing is currently not possible in public, due to limitations in subscriptions.
Some services and features cannot be tested properly, due to limitations in docker.
Grub is not configured by this role.

Testing

Most of the "generic" tests are located in the Test Library.

Ansible specific testing is done with Molecule.

Infrastructure testing is done with testinfra.

Automated testing is done with Travis CI.

Contribute

Thank you so much for considering to contribute. We are very happy, when somebody is joining the hard work. Please fell free to open Bugs, Feature Requests or Pull Requests after reading the Contribution Guideline.

See who has contributed already in the kudos.txt.

License

This work is licensed under a BSD-3-Clause License.

while-true-do/ansible-role-sys_kdump