Page 1 of 2
High Load Advice
Posted: Mon Feb 02, 2015 7:10 am
by btemple
We are reaching the end of our full implementation of NAgios XI. We seem to be having high load issues. I was hoping to get some advice. What stats hardware specs can I post for you to take a look ?
I am pretty sure our load spikes are coming from the netapp checks that we run. We have tons of volumes and the response time is a 1-2 seconds per cehck. The checks are currently run ever ten minutes.
Re: High Load Advice
Posted: Mon Feb 02, 2015 7:24 am
by btemple
More info. There was an obvious change on Jan 18th. I believe this is when the final Net app filer was added to production.
Re: High Load Advice
Posted: Mon Feb 02, 2015 9:42 am
by abrist
First, we need some info about the server hardware:
Code: Select all
lscpu > /tmp/info.txt
cat /proc/cpuinfo >> /tmp/info.txt
free -m >> /tmp/info.txt
top -n 1 -b >> /tmp/info.txt
ps -aef >> /tmp/info.txt
Post the file /tmp/info.txt.
Re: High Load Advice
Posted: Mon Feb 02, 2015 9:53 am
by btemple
Here you go thanks !
Re: High Load Advice
Posted: Mon Feb 02, 2015 11:21 am
by lmiltchev
Are you currently using modgearman or livestatus? Can you post the nagios.cfg file? Have you tried following any of the steps, outlined in the
"Maximizing Performance In Nagios XI" document?
Re: High Load Advice
Posted: Mon Feb 02, 2015 12:25 pm
by btemple
Not sure for question #1 where would I find the info ?
Answer # 2 attched
Will continue to work through #3
Re: High Load Advice
Posted: Mon Feb 02, 2015 12:56 pm
by scottwilkerson
Lets also run the following so we can see which process is using the CPU, I noticed tripwire on here which I know doesn't come on a default XI install
Re: High Load Advice
Posted: Mon Feb 02, 2015 1:21 pm
by btemple
See attached. Tripwire was installed due to separation of duties auditing
Re: High Load Advice
Posted: Mon Feb 02, 2015 4:46 pm
by scottwilkerson
Are you still seeing similar high load?
One thing I noticed in your last post was that the check_netapp_ng_64 plugin was using a considerable amount of CPU (53.5%).
We didn't create this plugin, nor do we have a netapp server to test against, but it certainly could be a bug in the plugin.
Re: High Load Advice
Posted: Tue Feb 03, 2015 7:04 am
by btemple
Load comes and goes. I would say it is still high. I originally posted that check as a possible root cause so I am pretty sure that is the issue.