This issue has been migrated from Redmine: https://dev.icinga.com/issues/8425

Created by mfriedrich on 2015-02-12 12:06:07 00:00

Assignee: mfriedrich
Status: Resolved (closed on 2015-02-12 12:25:05 00:00)
Target Version: 2.3.0
Last Update: 2015-02-12 12:25:05 00:00 (in Redmine)

Icinga Version: 2.2.4
Backport?: Not yet backported

There are two DbEvents::AddDowntime() events triggered asynchronously:

• DbHostObject::OnConfigUpdate() calls AddDowntimes() which adds all scheduled and configured downtimes
• OnDowntimeAdded signal gets triggered from checkable-downtime.cpp for each newly created downtime at runtime (which happens for the first downtime window for ScheduledDowntime config objects)

Since these events happen independant one from another, we cannot control whether one or the other should not insert such a downtime. And it's also not clear if an update would match on an insert before.

Therefore the solution is to delete the existing downtime before adding it again. This keeps the inserts intact, and ensures that there are not duplicate entries inside this status table.

This problem happens separately from #7765 where a config table with multiple config objects is involved.

Changesets

2015-02-12 12:19:27 00:00 by (unknown) 48fa1a1

DB IDO: Fix duplicate entries in scheduleddowntime table

fixes #8425

2015-02-12 13:22:24 00:00 by (unknown) 1c4501d

DB IDO: Require schema version 1.13.0

refs #8425

Relations:

• relates #8425