Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I monitor desktops during work hours. Some desktops do not get turned only once or twice a week. I have disabled host notifications for those, but active checks still fail and "host down" still shows up as an unhandled problem in nagios/multisite. How to make host down to not be a problem? Also I want active checks to stay "OK" (last state) while host is down.

@Posipiet: all of our desktops run linux (debian or ubuntu) and I want to make sure that they are connected to internet, no updates are pending, not running out of disk space and memory etc. Some of our staff are not technical enough to administer machines on their own. I want to get alert, ssh in, and fix it without users noticing. Yet I don't want to be spammed when they are on holiday.

share|improve this question
2  
hands down best post title ever! –  pfo Aug 24 '11 at 8:46
1  
What are you trying to achieve with Nagios? Please describe what problems you wish to solve. Perhaps some other tools fits your task better? –  Posipiet Aug 24 '11 at 8:48
    
Out of curiosity, why are you nagios monitoring desktops at all ? –  Sirex Aug 24 '11 at 8:55
    
Change the "host_check" command to "true" in the appropriate service. My advice may sound weird but fits to you question. –  SamK Aug 24 '11 at 8:57
    
@SamKrieg: That won't work because the service checks will go bananas. –  womble Aug 24 '11 at 8:59
show 4 more comments

4 Answers

I dont think, that desktop pcs should be monitored by Nagios, but maybe Allow host down while under nagios watch can help you.

share|improve this answer
1  
I agree, desktops under Nagios isn't something I'd ever do. –  womble Aug 24 '11 at 9:00
add comment

Perhaps you should look into an inventory software like OCSinventory: http://www.ocsinventory-ng.org/en/

Nagios is not meant to monitor systems that are not always on. Making it somewhat work in this situation will always be a kludge.

For workstations, we use OCSinventory, which also lists disk space etc. For servers, we use Icinga.

share|improve this answer
    
Correction: Nagios isn't meant to monitor systems that aren't reliably on. It'll handle a machine that's only up from (say) 9am to 5pm Monday to Friday, but it doesn't cope with machines where a human can't tell it in advance when it's going to be on or off. –  womble Aug 24 '11 at 22:06
add comment

Scheduled Downtime maybe is what you want.

share|improve this answer
add comment

Whilst I really don't like the idea of monitoring inherently unreliable hosts in Nagios, I have a possible solution for you. Implement at your own risk.

Install a script on every desktop that will tell Nagios "I'm going away now, and it's all OK" when the machine shuts down cleanly, and conversely says "I'm back, please monitor me" when it boots. If you make it a request to a CGI on the client side, you can define whatever mechanism you like on the server-side at your leisure. I'd probably have that script disable all host and service checks entirely, but you could have it set (a really long) downtime, or whatever took your fancy.

This doesn't work if people have a habit of hard-cycling their machines, but that's a user training issue (because you don't want them doing that even without a Nagios quell script).

Otherwise, given that you really only care about a few incidental non-services (pending updates, disk space, etc), I'd probably just ditch active checks entirely and go with passive checks. These are more like SNMP traps than active monitoring, and they fit the Nagios model much better.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.