2012R1.0 needs fix for "Undefined index: uid"

[paul.jobb]

Yes I removed the retention.dat file and restarted nagios, it didn't appear to make much difference with the nagios start up times.

[mguthrie]

Do you utilize a lot of recurring downtime in your environment?

Our CCM is a forked version of NagiosQL as of 3.0, so the NagiosQL changelog is no longer applicable for Nagios XI.

[paul.jobb]

No we haven't used any recurring down time.

We upgraded to 1.4 this morning which seems to have lowered cpu utilization on our mysql server(offloaded to separate server) noticeably , it was averaging 35% which has appeared to drop to 20%. I'm thinking that may have an impact on the restart time as certain parts of the interface definitely seem more responsive.

There was another thread that indicated in certain instances, nagios will start before ndo2b on system boot, that's definitely the case for this server. I'm not sure its related but thought I would mention it.

http://support.nagios.com/forum/viewtop ... lue#p39330

[mguthrie]

Ok, so lets see if we can pin this down just a bit more. Is the issue with the time it takes to Apply Configuration, or with running:

Code: Select all

service nagios restart

If your hosts number in the thousands, the Apply Configuration is going to take a bit to write new configs particularly for all of the hosts. We looked at having XI only write configs that were updated, but we had to halt that idea for a bit because things broke badly if an object got renamed.

[paul.jobb]

its definitely the writing of the configuration file, from the time the first host record is written(3:21:33) to the last service record is wrote(3:28:19) its about 7 minutes. Nagios restarts on its own fairly fast the first 3 checkmarks in the top right hand corner are green at start and the other three checkmarks stay blue for a few minutes and then turn green, however monitoring is occurring during that time.

[slansing]

I can speak to the issue you are seeing with the check marks, on medium to large installations Nagios starts itself, then the databases, and afterwords starts the active host/service/notification crons. It can take some time for those to show up when you spool the server or Nagios back to its normal state. As for decreasing the amount of time between config writes you can increase the memory limit in php to 512 and that should help a bit. We are looking at ways to change this for a future release but it will take quite a bit of testing to make sure that it does not cripple an installation if it breaks.

To change the memory limit run the following commands:

Code: Select all

vi /etc/php.ini

edit this line with the value included: memory_limit = 512M