Nagios Support Forum

Hey guys,

After our 2012 upgrade we've run into one major issue and two small issue, here's the list in order of criticality:

1. Nagios XI failing to load host/service information... this one is interesting. You can open the NOC screen just fine or network outages or any page that uses host or service information... but if you try to open any of the following pages it will just sit there with the spinning ajax ring of death:
- Open Host/Service problems
- All Host/Service problems
- Quick find host/service
- Host/Service detail

Everything else from a loading information perspective appears fine. This issue didn't show up initially after deployment and only surfaced some time this morning. This ones a bit of a major issue as users can't silence alerts or look up info on a host, only admins can via the Nagios Core interface.

2. Errors in /var/log/messages on Nagios restarts triggered from I think the legacy CCM restarts (I can't test this during business hours so it will have to wait until out of hours to work out the root cause).
Nov 5 11:54:33 nxi ndo2db: Error: mysql_query() failed for 'UPDATE nagios_conninfo SET disconnect_time=NOW(), last_checkin_time=NOW(), data_end_time=FROM_UNIXTIME(0), bytes_processed='0', lines_processed='0', entries_processed='0' WHERE conninfo_id='0''
Nov 5 11:54:33 nxi ndo2db: mysql_error: 'MySQL server has gone away'
Nov 5 11:54:33 nxi ndo2db: Error: Connection to MySQL database has been lost!

3. Capacity planning will load a graph when you select the page or select a host but as soon as you try to calculate future growth it will never load.

#1. We updated the Jquery libraries in 1.1 to fix some issues in IE. The browser is still probably caching the old versions though. Rick-click on the page and select "reload" and it should grab the fresh files. We've got a fix coming in 1.2 where updates like that won't cause anymore caching issues.

#2. We'll keep an eye out for this one. Even though the CCM is new, the Apply Configuration process still utilizes the same method as the 2011 version.

#3. Check for a "segmentation fault" error in the apache logs. We've seen this on a couple of systems, and it appears to be related to the sourcegaurdian encryption. We're posting a fix for this in the 1.2 release (hopefully today).

#1. Yep, this is definitely the issue, mystery solved

#3. Spot on again:
[Tue Nov 06 22:02:05 2012] [error] [client 172.31.121.248] PHP Fatal error: Allowed memory size of 67108864 bytes exhausted (tried to allocate 10223630 bytes) in /usr/local/nagiosxi/html/includes/utils-reports.inc.php on line 449, referer: http://kapnxi001/nagiosxi/reports/
[Tue Nov 06 22:02:50 2012] [notice] child pid 17758 exit signal Segmentation fault (11)
[Tue Nov 06 22:02:50 2012] [notice] child pid 15682 exit signal Segmentation fault (11)

I noticed the 1.2 source installer is up so I will get that in as soon as possible. Cheers Mike.

For error #2, check your connection limit for SQL and make sure you're not hitting the max on it.

Definitely not hitting the connection limit... never even reached 50 connections and the max is 200. It also happens exactly when "restart nagios" is pressed in both the old and new CCM, I think ndoutils was updated when we upgraded to 2012... perhaps an issue with the new version of that?

jsmurphy wrote:Definitely not hitting the connection limit... never even reached 50 connections and the max is 200. It also happens exactly when "restart nagios" is pressed in both the old and new CCM, I think ndoutils was updated when we upgraded to 2012... perhaps an issue with the new version of that?

hi js,

We also had an issue quite similar in nature when this happens can you check what top shows for mysql when this happens. do you see any performance issues when generating notification reports ?

Many thanks

arnab

arnab.roy wrote: We also had an issue quite similar in nature when this happens can you check what top shows for mysql when this happens. do you see any performance issues when generating notification reports ?

Hey Arnab,
Yesterday when working with some users to show them how to generate availability reports I noticed it seemed to take much longer in IE. But it was IE7 or 8 and pre-9 IE's JS engine ran slower than a three legged dog.... so I didn't really think much of it. Also as far as I know it could have always taken much longer in IE :p

The SQL error doesn't seem to do any harm or cause any performance issues so far as I can tell.