"No performance graphs were found for this host."

[jladd]

Well just as an update. I have since completely restarted on my nagios XI box project. I had some better luck installing this time around but I am still having the same issues I don't know if this is directly related to how the switch is reporting the data or what. This time on setting things up i set the kbps option instead of mbps and it didn't seem to change anything I don't have the UB units anymore but i do have things registering in b/sec instead of KB or MB for example 4096b/sec instead of 4kb/sec. But the graphs are spot on correct.

[mmestnik]

My guess is that if the number is less then 5 or 7kb for example that it auto converts units like a multimeter would. This could be seen as a feature, unless it's still reporting bytes after say 10 or 20kb.

[pbernardi]

Also, in the 'Service Detail' view, the original entry created with the config wiz, which contains cpu,mem,dis,etc and a couple of services shows up in caps. Additional services that I have added (via nagios core) show up under the server name with lower case letters.
I have attached a couple of screenshots... dc1bcx03 is the server in question.

nagios2.jpg

nagios1.jpg

[mmestnik]

I got some more information on this. Nagios Core(and thus NagiosXI) should treat host/services with different case as different hosts, just like UNIX treats file with different case. There is a bug in the glue used by 1.1H to communicate with Nagios Core. This glue was updated in the current development release and will be part of the next release 1.2.

http://go.nagios.com/tracker/40

This ha been corrected in the current Development release.

[pbernardi]

So, when I upgrade to the next version this will automatically be fixed? Or will I have to upgrade, delete the host, and then re-add it?
Do you also have an estimate as to when the release date is?

Thanks for looking into this.

Cheers,
Paul

[tonyyarusso]

So, when I upgrade to the next version this will automatically be fixed? Or will I have to upgrade, delete the host, and then re-add it?

Maybe. We of course hope it will, but that will depend on what comes up in testing.

Do you also have an estimate as to when the release date is?

That depends entirely on how the testing process goes and what kind of feedback we get. If you can help out with that, it will speed the release up. You can download a development snapshot from http://assets.nagios.com/downloads/nagi ... vel.tar.gz and give that a try, and report how it went for you.

[dxf1]

Hi Config resolved at core level, some issues with a group and two servers deleted in GUI, revisited the core and resolved issues readded the missing group and services all fine thanks

[Box293]

Tony,
I experienced this problem as well. The 1.2 pre-release fixed part of my problem.

History:
I originally typed my hosts in caps when creating them using the wizard.
Decieded to change it later and I had the issues as shown above in the screenshots.
I had also deleted hosts / services and re-added with wizard but same problem.
The 1.2 pre release fixed this. All my services display correctly under one host name, this host name is the correct one.

However the performance graphs for the service did not start working, they were still missing after 20 minutes.

In the folder /usr/local/nagios/share/perfdata/hostname/ I deleted the .rrd and .xml files for the service.
After doing this all came good again in about 5 minutes.

One thing I noticed is that after deleting the files for 8 services, if I went and looked at all 8 services to see if a performance graph existed, the Nagios XI server was much slower. Not always but sometimes the Performance Grapher component will stop and will need a manual restart. Looking at the CPU usage of the Nagios XI server you can see there is a heafty increase of CPU usage as compared to previous trends.

[mmestnik]

So you believe that PNP may also have an issue with case changes?

The behavior we are looking for at this time is a delete/create. Every thing should behave as thought the names are completely different, even if it is something as simple as a change in case.

In the future we may add a tool to rename hosts or re-describe services, but currently there are no plans to do so. One thing that I believe is that this is going to be more popular on new installations then on systems that have been running for a long time. An admin could decide that he doesn't like his hostnames, but I believe this will be rare and cause more headaches then the loss of historical data.

[Box293]

So you believe that PNP may also have an issue with case changes?

Yes, it seems that it needs old files deleted for it to start displaying new data (graphs).

To provide more history:

• I originally typed my hosts in caps/lowercase when creating them using the wizard
  Because this was an ESX host I needed to type in the FQDN not the IP address
  Something like SERVERNAME.DOMAIN.local
  I also called the friendly hostname the same
  I then realised that the VMware SDK API was case sensative
  So I changed the address of the host to servername.domain.local
  But I left the friendly hostname SERVERNAME.DOMAIN.local
  This fixed my immediate problem and performance graphs worked OK
  After a week or so I decided that all hosts would be in lowercase
  I changed my hosts and I think at this stage the graphs did not work any more.
  I think then I deleted *some* of the .rrd and .xml files
  I lost all existing data but graphs started to be generated again
  I did not get around to deleting all the old files
  I also by this stage, from what I remember, this is about when the issue appeared with services appearing in the Service Detail with different case, just like the screenshots in this topic.
  After this it's as I've already described in my previous post (after upgrading to 1.2 dev release)

we may add a tool ... One thing that I believe is that this is going to be more popular on new installations then on systems that have been running for a long time.

I do agree with you here.

I think once people play with Nagios XI for a short while they soon work out what their naming standard is going to be. It's probably going to be based on something existing anyway. So it's unlikely things are going to change and it would cause more headached than it's worth.

It's more likely that someone sets up Nagios XI very quickly without following any naming standards. They later change things and realise they've lost some performance data. I think your tool would come in very handy here.