[dev.icinga.com #11590] notification interval = 0 not honoured in HA clusters #4121
Comments
Updated by mfriedrich on 2016-04-14 09:24:34 +00:00
|
Updated by mfriedrich on 2016-04-14 09:25:18 +00:00
This is most likely a duplicate of #11562. |
Updated by mfriedrich on 2016-05-06 13:38:06 +00:00
|
Updated by mfriedrich on 2016-05-06 13:38:36 +00:00
|
Updated by mfriedrich on 2016-06-07 19:14:47 +00:00
Left node feels responsible for sending the notification first, then you'll restart/stop the left node. Then the right node is responsible for the notification object. It does not get the last_problem_notification timestamp synced (#11562) and will therefore fire its first local notification. The user gets 2 notifications and thinks that interval = 0 does not work. Apparently it is the missing cluster exchange of notification messages - each node normally calculates the notification times on their own. |
Updated by mfriedrich on 2016-08-08 11:17:30 +00:00
|
Updated by mfriedrich on 2016-08-16 12:47:17 +00:00
|
This issue has been migrated from Redmine: https://dev.icinga.com/issues/11590
Created by mnardin on 2016-04-14 09:13:01 +00:00
Assignee: gbeutner
Status: Closed (closed on 2016-06-15 16:09:31 +00:00)
Target Version: 2.5.0
Last Update: 2016-08-08 11:17:30 +00:00 (in Redmine)
Hi,
in our setup (2 masters, both with notification features enabled) a notification object with interval = 0 will send out more than one notification:
debug.log on master1:
debug.log on master2:
I don't understand why this happens even with interval = 0:
Regards
Mirko
Relations:
The text was updated successfully, but these errors were encountered: