NDOutils and Nagios startup speed

[Mathieu Gagné]

Hi,

Marc Powell wrote:
> On Jan 26, 2009, at 4:46 PM, Mathieu Gagné wrote:
> 
>> Unfortunately, we can't afford this kind of downtime while
>> Nagios/NDOutils is busy exporting to MySQL. Also, host/service status
>> are not available while the reload is occuring.
> 
> While I can't really speak to the more general problem (I experience  
> it too), I did find the 'Retain status file over a reload' patch at http://altinity.blogs.com/dotorg/2007/09/nagios-patch-da.html 
>   to be extremely helpful for this particular symptom. Reloads are  
> completely transparent to GUI users now.

I never had this issue when sending the external command 
"RESTART_PROGRAM" to Nagios. Host/service status are still available to me.

My understanding is that "RESTART_PROGRAM" is the equivalent of issuing 
a "reload" or sending a HUP signal. Unfortunately, during that time, 
checks are still not performed as Nagios is busy broking events to 
NDOutils and to MySQL in a synchronous way.

I found this patch on the same page:
Do not resend retained status to NDO
http://svn.opsview.org/opsview/tags/nagios-patch-day/opsview-base/patches/nagios_stop_logging_retained_states_to_ndo.patch

The following questions come to my mind:

- What could be the consequences of using such patch?

- When will old entries from the table "nagios_hoststatus" be deleted? 
(same for "nagios_servicestatus")

- What does aggregated_dump mean? (the patch changes it from FALSE to TRUE)
-> int update_host_status(host *hst,int aggregated_dump){
-> int update_service_status(service *svc,int aggregated_dump){

- The comment tells: "a check cycle needs to complete before NDO has all 
the status information". What does it mean? To which "check cycle" does 
it refer to?

- Does it have any link to "aggregate_status_updates" and 
"status_update_interval"? Does it mean status will be dump later and not 
at startup? "status_update_interval" seconds after startup?


>> Is there a way to speed things up? Any help would be appreciated.  
>> Thanks.
> 
> I'm interested in tips as well.

I had to create several SQL indexes. At startup, the broker deletes old 
entries and limits the deletion by "start_time" and/or "scheduled_time". 
Unfortunately, there is no indexes on tables with such columns, putting 
MySQL on its knees as soon as Nagios restarts.

Disclaimer: I'm not a DBA and can't guaranty the efficiency of each of 
them. I might have also missed some.

ALTER TABLE `nagios_servicechecks` ADD INDEX `start_time` ( 
`instance_id` , `start_time` )  ;
ALTER TABLE `nagios_timedevents` ADD INDEX `scheduled_time` ( 
`instance_id` , `scheduled_time` );
ALTER TABLE `nagios_systemcommands` ADD INDEX `start_time` ( 
`instance_id` , `start_time` );
ALTER TABLE `nagios_hostchecks` ADD INDEX `start_time` ( `instance_id` , 
`start_time` );
ALTER TABLE `nagios_eventhandlers` ADD INDEX `start_time` ( 
`instance_id` , `start_time` );
ALTER TABLE `nagios_instances` ADD INDEX `instance_name` ( 
`instance_name`,`instance_id` )  ;
ALTER TABLE `nagios_hosts` ADD INDEX host_object_id (host_object_id);


--
Mathieu

------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null