Change in ranges for ScheduledDowntime is not applied.

[StoneISStephan]

icinga2 2.6.3-1ppa1xenial1 r2.6.3-1

When changing the ranges in a Host object the downtimes already scheduled in the database are not adjusted and neither are the new downtime schedules created.

icinga2 object list --name 'hostname.tld!http!update-downtime-service'

Object 'hostname.tld!http!update-downtime-service' of type 'ScheduledDowntime':
% declared in '/etc/icinga2/conf.d/downtime.conf', lines 28:1-28:60

• __name = "hostname.tld!http!update-downtime-service"
• author = "icingaadmin"
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 29:2-29:23
• comment = "Scheduled downtime for System Updates Applied to Services"
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 30:2-30:70
• duration = 0
• fixed = true
• host_name = "hostname.tld"
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 28:1-28:60
• name = "update-downtime-service"
• package = "_etc"
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 28:1-28:60
• ranges
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 32:2-32:42
  • wednesday = "01:00-02:00"
• service_name = "http"
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 28:1-28:60
• templates = [ "update-downtime-service" ]
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 28:1-28:60
• type = "ScheduledDowntime"
• vars = null
• zone = "XXX"
  % = modified in '/etc/icinga2/conf.d/downtime.conf', lines 28:1-28:60

After which a systemctl restart icinga2 was issued.

But after 10 minutes the DB still shows the old range (from 04.00 till 05.00):

mysql> SELECT * FROM icinga_scheduleddowntime WHERE name LIKE 'hostname.tld%http%' \G
*************************** 1. row ***************************
scheduleddowntime_id: 3631
instance_id: 1
downtime_type: 1
object_id: 5383
entry_time: 2017-06-14 11:21:16
author_name: icingaadmin
comment_data: Scheduled downtime for System Updates Applied to Services
internal_downtime_id: 10
triggered_by_id: 0
is_fixed: 1
duration: 0
scheduled_start_time: 2017-06-21 04:00:00
scheduled_end_time: 2017-06-21 05:00:00
was_started: 0
actual_start_time: 2017-06-21 04:00:00
actual_start_time_usec: 0
is_in_effect: 0
trigger_time: NULL
name: hostname.tld!http!stone-mon001-1497432076-10
endpoint_object_id: 325
session_token: 1497435587
1 row in set (0.00 sec)

And this remains even hours after changing the config and restarting icinga2.

This ultimately causes us to be unable to update/adjust any previously set ScheduledDowntime's.

The only workaround here is to remove all:

vars.update_downtime = {
	ranges = {
		"wednesday" = "04:00-05:00"
	}

from ALL Host Objects, which is no solution,
and reload/restart Icinga2, which deletes the currently scheduled downtimes.

And afterwards add again:

vars.update_downtime = {
	ranges = {
		"wednesday" = "01:00-03:00"
	}

and reload/restart Icinga 2 again, which creates new scheduled downtimes.

[dnsmichi]

Could be related to #4790.

[njm506]

It's not just the ranges; it looks like nothing gets updated as long as the original configuration object exists. If you change the name or delete the configuration object it gets corrected, though.

[dnsmichi]

Should be fixed with 2.10.3.