PDA

View Full Version : problem setting up maintenance


ulukay
15-02-2010, 14:09
hi

i'm trying to setup a weekly maintenance for windows servers, so i don't get flooded with emails. i'm not sure why, but i got some problems there :o

http://666kb.com/i/bgq3bag9olqrgbao2.png
this should set all windows servers into maintenance mode every sunday 20:00pm for 10 hours

and this action should send me emails only if a windows server got a problem AND is not in maintenance mode
http://666kb.com/i/bgq3b8pb3c4swgszm.png

unfortunately this isn't working and i receive massive amounts of emails every sunday/monday :(

ulukay
19-02-2010, 06:46
*bump*

it's getting weekend again :eek:

arli
19-02-2010, 08:31
I tested a setup like yours on 1.8.1 and it worked out just fine. Are you using a 1.8 or 1.8.1 version?

ulukay
19-02-2010, 09:13
i'm using 1.8.1 too :(

arli
19-02-2010, 09:39
If you don't get any better advice before weekend, then maybe you could narrow the problem down, by creating separate action for only windows servers group, to eliminate the possible error in the trigger condition calculation?

ulukay
19-02-2010, 09:51
that's a good idea, i'll try that!

ulukay
22-02-2010, 06:38
well, didn't work either :confused:

simonc
24-02-2010, 14:59
did u resolv your problem ?
i'll try that this night ;)

ulukay
24-02-2010, 15:07
nope, still getting spammed :(

mizaring
26-02-2010, 12:45
Hi all,

I'm getting the same problem as ulukay. I have some PDU/UPS that do internal testing every monday evening. So I set up a weekly maintenance each monday between 17h and 22h. However when I come back tuesday morning I get over 2000+ emails from zabbix. The emails are only the "recovery" messages, not the "problem" ones. Looks like zabbix get stuck on something and spam the s**t out of my mailbox! I can reproduce the problem easily by configuring a trigger, than a maintenance period with that host and finaly an action with recovery message. Now when we the trigger come on in the maintenance period zabbix start to spam with recovery message. Any other out there with the same problem?

f.koch
27-02-2010, 18:44
hi @all

I have the same behavior like ulukay.
I try setting a maintenance with on time only or daily/weekly I was always spamed with mails (not only recovery, also problem).

Zabbix Server Version 1.8.1
Zabbix Agent Version 1.8.1

mizaring
02-03-2010, 16:53
Hi,

Still got spammed this morning... only this time with 108 emails instead of over 2000. This time I got all "problem" emails, not "recovery" ones. However it's still very problematic since we have an automated process that open a support ticket each time an email is sent, imagine trying to close 2000+ tickets... We also have support techs that would be called if I didn't remove the usual process from our UPS. Could a developer get a look on this? Just email me if you need more information on my configuration.

Thanks,

arli
04-03-2010, 09:20
The emails are only the "recovery" messages, not the "problem" ones.

Are you using normal event generation, or normal + multiple true events?

mizaring
04-03-2010, 15:10
Are you using normal event generation, or normal + multiple true events?

Normal event generation, not "normal + multiple true events". I checked that in the beginning, just to be sure! Note that last monday I received only the "problem" mails, not the "recovery" ones.

Maybe f.koch and ulukay can confirme that the problem is not related to the event generation.

Thanks!

f.koch
04-03-2010, 20:59
Normal event generation, not "normal + multiple true events". I checked that in the beginning, just to be sure! Note that last monday I received only the "problem" mails, not the "recovery" ones.

Maybe f.koch and ulukay can confirme that the problem is not related to the event generation.

Thanks!

I don't know what you mean with "multiple true events"?

mizaring
04-03-2010, 21:19
I don't know what you mean with "multiple true events"?

When you create a trigger you have an option called "Event generation" with 2 possible value :

- Normal – events are generated normally, on TRIGGER status change
- Normal + Multiple TRUE events – events are also generated on every TRUE evaluation of the trigger

http://www.zabbix.com/documentation/1.8/manual/web_interface/configuration

Just make sure your triggers are "Normal" and not "Normal + Multiple TRUE events"

f.koch
05-03-2010, 07:09
When you create a trigger you have an option called "Event generation" with 2 possible value :

- Normal – events are generated normally, on TRIGGER status change
- Normal + Multiple TRUE events – events are also generated on every TRUE evaluation of the trigger

http://www.zabbix.com/documentation/1.8/manual/web_interface/configuration

Just make sure your triggers are "Normal" and not "Normal + Multiple TRUE events"

ok thx, I use "normal" event generation.

mizaring
10-03-2010, 20:39
Hi all,

I didn't receive any email this week... but I restarted the zabbix-server last week. So I guess maybe the modifications in the database are not picked up correctly by the server until you reboot it. I also found a way to reproduce the problem rather easily :

1 - Create items, triggers and actions for a host
2 - Activate one of the triggers that send a mail (you should receive an email activated by the trigger)
3 - Enter a maintenance period from now till some time in the future while the trigger is still active.
4 - Wait for 10 minutes... then try to toggle the trigger (on/off).

Works almost all the time... but be prepared to get spammed!

Please follow up on this problem!

Thanks!

Fluze
30-08-2010, 23:34
I also have the same problem.
Any news on the issue?
Zabbix 1.8.2 and 1.8.3.

mizaring
31-08-2010, 01:05
Hi all,

We traced the problem with the zabbix support to be a configuration problem with the action. When we checked "recovery message" in the action we thought it was related to the state of the trigger (Problem = send message, Ok = send recovery message). However the "recovery message" is sent when a trigger happen and the action conditions goes to "false". So if you don't use Trigger Value = Problem as a condition in your action you can end up with this problem because the action conditions may stay true even if the trigger value is back to Ok.

Hope that helps

Fluze
31-08-2010, 08:12
Thank you for the quick reply! i will try it and report back.

red1157
23-12-2010, 16:30
Same problem.... :(
Any news on the issue?
Zabbix agent 1.8.3
Zabbix server 1.8.3
CentOS 5.5.

red1157
23-12-2010, 17:32
Add condition "Trigger value=Problem" not work :(

red1157
23-12-2010, 17:36
And finally, if in action add condition "Maintenance status in "maintenance" - in my case zabbix NOT send any alerts then maintenance period active | not active and trigger false | true :(
I do not understand conditions "Maintenance status in "maintenance"....... :confused:

arli
23-12-2010, 19:01
Same problem.... :(
Any news on the issue?
Zabbix agent 1.8.3
Zabbix server 1.8.3
CentOS 5.5.

This issue described here earlier is definitely fixed in 1.8.3

red1157
27-12-2010, 14:30
I do not see: ( tell the post № pls

arli
27-12-2010, 15:45
Sorry... I posted this reply on a wrong thread. I never managed to reproduce the problem on my environment.

scanch
14-01-2011, 11:21
Up !
Same problem here when migrating from 1.8.3 to 1.8.4.
Without changing configuration, in the 1.8.3 version we don't receive email alerts when host are in maintenance mode, but now we are spammed. :(

Example that we have now in 1.8.4:
- Action is to send mail when maintenance status IS NOT in maintenance.
- Maintenance period all days at 23h50 with duration of 1h40.
- Database stopped a 0h00, zabbix send mail at 0h01 about the state of the database, subject start with "RECOVERY: ..... PROBLEM".
- From the Event tab I can see trigger is OK at 0h47, no mail are sent.
- At the end of the maintenance period (1:30AM), one minute later, Zabbix send mails about the state of the database: "Alert: ..... OK".

mjsr
10-05-2011, 11:06
Hi there,

Just noticed *that* specific weird behaviour on our 1.8.4 installation.

Just after the maintenance period finishes, we get a notification that the host is OK.

In the worse case it could tell "{HOSTNAME} finished maintenance period 'database shutdown for full backup'" or whatever.

Thoughts? Suggestions? Ideas?

tyvm!

igor
23-05-2011, 12:53
Zabbix is working correctly in this case.

In order to get rid of these OK notifications after the maintenance ends you need to perform the following changes in the configuration of your action:

1. In the configuration of your action from "Action conditions" you should remove option "trigger value = "OK"" and leave option "Trigger value = "PROBLEM"" and also leave option "Maintenance status not in "maintenance"";

2. Turn on (enable) recovery messages in this action;

This will allow you NOT to receive notifications about the Problem events that occurred during the Maintenance period and also will not send recovery messages (OK messages) when maintenance will ends.

untergeek
23-05-2011, 17:21
Please post the full details of both your maintenance window as well as your notification Action. These details are necessary to fully troubleshoot. You can omit identifying details, but otherwise we need all of the other details, down to the last checkbox.