Updated by mfriedrich on 2015-12-18 17:17:55 00:00

• Category set to DB IDO
• Target Version set to 2.4.2

Updated by elippmann on 2015-12-21 11:51:00 00:00

I can confirm this.

Updated by mfriedrich on 2015-12-22 10:58:36 00:00

• Priority changed from High to Urgent

We'll look into that after our xmas holidays (11.1.2016).

Updated by mthode on 2015-12-24 04:32:26 00:00

Do you have more info on this? I think I'm hitting it, but don't know.

Updated by tgelf on 2015-12-24 08:53:18 00:00

Unfortunately I guess because of Christmas holidays there will not happen much on this issue within the next weeks. I tried to work around this problem with a cronjob running once a minute doing as follows:

UPDATE icinga_objects o join icinga_hoststatus s ON s.host_object_id =
o.object_id set o.is_active = 1 where o.objecttype_id = 1 and
o.is_active = 0 AND s.last_check > NOW() - INTERVAL 1 MINUTE;
UPDATE icinga_objects o join icinga_servicestatus s ON
s.service_object_id = o.object_id set o.is_active = 1 where
o.objecttype_id = 1 and o.is_active = 0 AND s.last_check > NOW() -
INTERVAL 1 MINUTE;

Only drawback: in theory it could happen that this re-enables a deleted object in IDO, let me know when this happens to you - I could try to add a few more checks to the query.

Cheers,
Thomas

Updated by mfriedrich on 2015-12-24 09:55:03 00:00

This problem only exists in the current git master and does not affect stable releases.

As said, we are going to look into it after the holiday season enjoying our offline time for a while.

Happy holidays!

Michael

Updated by tgelf on 2015-12-24 10:09:32 00:00

@mthode: Be assured that Michi and Gunnar absolutely deserve their holidays :-) I'd personally still suggest to stay with the latest snapshot if the above queries help to mitigate your problem. They did a very good job in fixing a lot of severe issues in 2.4.x, so staying with 2.4.1 was not an option for me.

Cheers,
Thomas

Updated by gbeutner on 2016-01-11 13:24:57 00:00

• Status changed from New to Assigned
• Assigned to set to gbeutner

Updated by Anonymous on 2016-01-11 13:25:02 00:00

• Status changed from Assigned to Resolved
• Done % changed from 0 to 100

Applied in changeset 1bf2701.

Updated by mfriedrich on 2016-01-11 13:49:16 00:00

Happens randomly, and returns a non-predictive number of active/inactive objects (http://localhost/icingaweb2/monitoring/list/services). The race condition happens inside the DbType::GetOrCreateByObjectName() not properly setting the object reference.

Verified working.

Updated by mfriedrich on 2016-02-04 16:13:13 00:00

• Relates set to 11065

Updated by gbeutner on 2016-02-23 09:58:35 00:00

• Backport? changed from Not yet backported to Already backported

Updated by dokon on 2016-03-16 08:57:05 00:00

Bug seems to be still there ... have a customer Setup where i get the following

----------~~------------------------------------------------------------------~~-----------

----------~~------------------------------------------------------------------~~-----------

----------~~------------------------------------------------------------------~~-----------

And the host is missing in the icingaweb2 host view , but viewable in icingaclassic... :-/

Updated by gbeutner on 2016-03-16 08:58:07 00:00

Please submit a new ticket for this.