nrhodes91 / auter-manager

Tools to help deploy Auter host based automatic updates

auter-manager

Presentation

This repo contains ansible playbooks to install and configure Auter on RHEL-derivative devices through automation using a single csv file to handle configuration settings.

The playbooks require a csv file to run (auter_config.csv by default). A template file auter_config_template.csv is provided as a sample, matching the following array in csv format:

NAME	EXCLUDES	AUTO_REBOOT	PKG_MGMT_OPTS	MAX_DELAY	FROM_PREP_ONLY	PREP_TIME	PREP_DAY	PREP_MONTH	APPLY_TIME	APPLY_DAY	APPLY_MONTH	ALWAYS_REBOOT
labtest1		yes		300	yes	01:00	1	1-11	00:10	monday:second	1-11
labtest2		yes		300	no	01:00	1		00:10	monday:second
labtest3	kernel*	no		300	yes	01:00	15	"1,3,5,7,9,11"	00:10	1	"2,4,6,8,10,12"	yes
labtest4	""	yes		300	yes	02:00	monday:first	"1-11"	01:00	monday:second	1-11	yes

Important

• Whenever setting a list of values, be mindful of wrapping the values with double quotes (" ") if using comma (,) inside a field (cf labtest3 in sample table).
• Setting a value of "" or '' for EXCLUDES field (ex: labtest4 above) will remove any existing excludes.

Things to consider when filling the csv file:

1. auter --prep and auter --apply can be set up to run at most once a month when using this playbook.
2. fields from PREPTIME onwards are only mandatory if you use playbook auter_scheduler.yml or the meta-playbook auter_manager.yml as they only concern the cron settings.
3. you cannot use cron notation */n to define periodicity, but ',' and '-' are supported for month setting.
4. PREPDAY and APPLYDAY can be entered either as a single number (1 to 28) or as a pair separated by colon : <lettered_day_of_the_week:nth_occurence> (nth_occurence among first, second, third or fourth)

• Valid examples:
  • monday:second
  • 10

Detailed description for each field

Field	Description	Default value
NAME	Ansible inventory hostname	N/A
EXCLUDES	List of packages to exclude	None
AUTO_REBOOT	Should the device be rebooted if packages have been updated?	yes
PKG_MGMT_OPTS	Options to pass to package manager. If dnf is installed, it will be the preferred package manager.	None
MAX_DELAY	Upper limit of a random time to wait before querying repositories. This applies to downloading updates (--prep) and installing updates (--apply). This is used to stagger load on the repository servers.	3600
FROM_PREP_ONLY	Only install from the packages downloaded during auter --prep.	no
PREP_TIME	Time to execute auter --prep	None
PREP_DAY	Day to execute auter --prep	None
PREP_MONTH	Month(s) to execute auter --prep	None
APPLY_TIME	Time to execute auter --apply	None
APPLY_DAY	Day to execute auter --apply	None
APPLY_MONTH	Month(s) to execute auter --apply	None
ALWAYS_REBOOT	Should the device systematically be rebooted after patching (even if no packages were updated)? This option superseeds AUTOREBOOT, and add the extra --reboot option to auter --apply in the /etc/cron.d/auter file.	no

Playbooks

• auter_manager.yml loads up auter_installer and auter_scheduler roles to install, configure and schedule Auter based on devices configured in the CSV file. It is intended to be the 'default' playbook to use to set up everything in one go.
• auter_install.yml will install and configure Auter based on devices configured in the CSV file.
• auter_scheduler.yml on its own handles the crontab configuration for the two main fonctions of Auter: prep and apply. The play will write the cron scheduling into file /etc/cron.d/auter
• auter_reporter.yml is an easy way to gather information on Auter status and latest use for any number of devices. The result file will be place into sub-directory output.

note: From v1.2 onwards, the auter_scheduler.yml play will pause and warn when deployed over a device which has auter either disabled or not installed. The play will carry on after 10 seconds anyway.

Why csv format?

Using the csv file should make it easy to keep configuration reference easily readable from a spreadsheet editor, so that it can be maintained by non-tech people.
note: you should not add/remove any columns or the playbook execution will break. The csv file should suffice as the only source for Auter installation / configuration / scheduling.