Re: [Nagios-devel] Removing host checks for non-OK passive results
Posted: Fri May 26, 2006 1:19 am
On Wed, 24 May 2006, Ton Voon wrote:
> On 24 May 2006, at 14:23, Bruce Campbell wrote:
>> On Wed, 24 May 2006, Ton Voon wrote:
>>> Going back to the original problem, would a retry_check_interval for host
>>> checks help with this particular case?
>>
>> I don't believe so. Nagios will still execute the host_check command each
>> time a service on that host has a non-OK state. Since you can have a dead
>> host which still responds to ping, this is a lot of executions of the
>> host_check command when you have a lot of services on said host.
>
> Sorry, I was unclear. I meant going back to Jan Kratochvil's original problem
> of the WAN link having outages of up to a minute. I think having a
> retry_check_interval would help there.
Assuming that the behaviour of the retry_check_interval was to quell any
attempt by Nagios to run the host_check command during that interval, it
would assist.
I suspect that the problem at heart is a difference between the
notification being generated after max_check_attempts executions of the
host_check command, and not after the intuitive max_check_attempts *
check_interval time. Since the host checks are run on-demand in the case
of a failing service, the max_check_attempts count goes down pretty
quickly.
--
Bruce Campbell
High Voltage finds a fool to be an excellent path to ground.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
> On 24 May 2006, at 14:23, Bruce Campbell wrote:
>> On Wed, 24 May 2006, Ton Voon wrote:
>>> Going back to the original problem, would a retry_check_interval for host
>>> checks help with this particular case?
>>
>> I don't believe so. Nagios will still execute the host_check command each
>> time a service on that host has a non-OK state. Since you can have a dead
>> host which still responds to ping, this is a lot of executions of the
>> host_check command when you have a lot of services on said host.
>
> Sorry, I was unclear. I meant going back to Jan Kratochvil's original problem
> of the WAN link having outages of up to a minute. I think having a
> retry_check_interval would help there.
Assuming that the behaviour of the retry_check_interval was to quell any
attempt by Nagios to run the host_check command during that interval, it
would assist.
I suspect that the problem at heart is a difference between the
notification being generated after max_check_attempts executions of the
host_check command, and not after the intuitive max_check_attempts *
check_interval time. Since the host checks are run on-demand in the case
of a failing service, the max_check_attempts count goes down pretty
quickly.
--
Bruce Campbell
High Voltage finds a fool to be an excellent path to ground.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]