"No performance graphs were found for this host."

[jladd]

I have got nagios xi setup and monitoring one of our HP switchs. Model 8000m. I set it up using the monitor wizard setup. It seems to me like some of the graphing and monitoring is off im not sure if its the switch or what. But for example port 1 which is a 10/100 mb port and this is my graph image.

To try and clarify what I am asking is the max in and out numbers. How could they be that high for a 10/100 port.

[jladd]

Something else i just noticed.



If you look all of the services have the same duration of uptime. even though some of the hosts were just added a few minutes ago not 1d21h6m27s ago.

Am i looking at this wrong or am i looking at some kind of an issue?

Has anyone else every reported this kind of issue or am i the only one alone in the dark here? Right now im still evaluating the xi product and would like to get it 100%.

Is this because im monitoring a hp switch possibly? is there a different plugin i could use out there that might work better can anyone give me some insight on this? As well when monitoring ports if someone shuts down a pc I get a warning in nagios telling me the port is down is there a way to make nagios stop monitoring the port when its down and then start monitoring it again when it comes back up?

[admin]

Just a quick note on these two issues:

1. The duration time you reported as abnormal was confirmed to be a bug. We'll have a fix for this out in the 1.1G release which is due Wednesday or Thursday of this week.

2. The bandwidth issue is strange. Can you post the plugin output that is shown in the service detail? That will contain the text string output that describes how much bandwidth is being used.

3. As part of the 1.1G release, we're changing the switch monitoring wizard to allow you to override the port names to a friendlier version if you'd like.

[jladd]

Is this what you were looking for?



Advanced Status Details
Service State: Ok
Duration: 1d 23h 22m 30s
State Type: Hard
Current Check: 1 of 5
Last Check: 2010-03-16 19:07:48
Next Check: 2010-03-16 19:10:48
Last State Change: 2010-03-14 19:46:48
Last Notification: Never
Check Type: Active
Check Latency: 0.015 seconds
Execution Time: 0.03658 seconds
State Change: 0%
Performance Data: in=.054667Mb/s;50;80 out=.176088Mb/s;50;80


I did apply the patch and it fixed a lot of things Thanks. But i do still have a few questions if thats ok. On the performance graph the numbers still appear to be off. 10/100mbps ports are showing upwards of 200+ mbps average rates Not all of them are doing this and some ports are showing measurments in ub/sec This is only in the performance graphs. It appears when you look at everything in the service detail view that its accurate. I was wondering what would be causing this? and is there something i can do like re-install the graphing piece of nagios to possibly fix this?

[mmestnik]

My best guess is this is a double correction from B to b, keep in mind counters are done by hosts in Bytes then * 8 to get bits for nice looking graphs.

This could be being done twice, would that sound about right to you?

We stopped using Cacti and switched to PNP in F.

It's still most likely a configuration issue. Can you verify the Performance data? If the graphs reflect this, then it's likely the problem is with the ??SNMP?? oids. One thing I just thought of was checking to make sure the Graph is not a *gauge.

* The number reported should be an ever increasing byte counter, you would then need to be graphing the change in this value over time and not the value it self.

We do understand that this feature is something lacking from our product. Are you able to run these commands from the command line? SSH(ing) into the server can make testing these things much simpler also these commands all have a '--help' argument.

[jladd]

Thats what i was thinking at first some kind of Byet's to Bits conversion. But looking at the max mb/sec on the last bandwidth graph would blow that out of the water especially it being a max of 100megabit per second. 22gbit a sec is way out of range. I was also wondering if anyone knows what is used by default by nagiosxi for the bandwidth graphing and reporting I still have some time for evaluation and was thinking about trying to see if i could get cacti going? or does XI use cacti already?

Ok is there anything you could recommend me to check or re-install that might fix this? and if i were to install cacti and get that working If i buy a service plain will it still be supported.

Is there any documentation you can show me so i can understand the configuration's of the plugins better. I started to look at them and play with them and noticed that i could clear up a few issues but im not experienced enough to be able to adjust all of the settings and what not?

I am sorry i am still new to nagios and still learning linux can you tell me which commands and how to run them from console?

[mmestnik]

It would look something like this:
http://markmail.org/message/unolu6ciep5ygekq

Code: Select all

root@Bungle:/usr/local/nagios/libexec# ./check_http -I pluto.uk.nsynergy.com -a
123:123
HTTP OK - HTTP/1.1 302 Redirect - 0.690 second response time
|time=0.690064s;;;0.000000 size=457B;;;0
root@Bungle:/usr/local/nagios/libexec# ./check_http -H pluto.uk.nsynergy.com -a
123:123
HTTP OK - HTTP/1.1 302 Redirect - 6.493 second response time
|time=6.493219s;;;0.000000 size=467B;;;0
root@Bungle:/usr/local/nagios/libexec# ./check_http -H pluto.uk.nsynergy.com -a
123:123
HTTP WARNING: HTTP/1.1 401 Unauthorized
root@Bungle:/usr/local/nagios/libexec#

With –vvv root@Bungle:/usr/local/nagios/libexec# ./check_http -H
pluto.uk.nsynergy.com -a 123:123 -vvv
GET / HTTP/1.0
User-Agent: check_http/v2053 (nagios-plugins 1.4.13)
Connection: close
Host: pluto.uk.nsynergy.com
Authorization: Basic bmFnaW9zOjB2M3IyWTB1

[jladd]

ok from the console while sshed in and doing the bandwidth check & status check on port 17 im showing

[root@localhost libexec]# ./check_rrdtraf -f /var/lib/mrtg/192.168.0.4_17.rrd -w 50,50 -c 80,80 -l M
OK - Current BW in: 0Mbps Out: 0Mbps|in=.000020Mb/s;50;80 out=.009779Mb/s;50;80
[root@localhost libexec]# ./check_ifoperstatus -H 192.168.0.4 -c public -k 17
OK: Interface C1 (index 17) is up.[root@localhost libexec]#

The advanced status details is showing

Service State: Ok
Duration: 23d 21h 1m 19s
State Type: Hard
Current Check: 1 of 5
Last Check: 2010-03-30 17:14:06
Next Check: 2010-03-30 17:17:06
Last State Change: 2010-03-06 19:15:25
Last Notification: Never
Check Type: Active
Check Latency: 0.228 seconds
Execution Time: 0.03792 seconds
State Change: 0%
Performance Data: in=.000020Mb/s;50;80 out=.009779Mb/s;50;80

So it appears to be ok but looking at the graph for port 17 i get this



The graph itself appears correct and looks right to me but it lists the measurements at the bottom and there being displayed in ub/sec instead of mb/sec. Thanks for all the help so far.

Not all of the ports appear this way Most of them are working great It just seems like some of the units of measurement are incorrect or not displaying properly.

[jladd]

Here is another example i found Is port 21 which on our switch is card C port 5.

Running the check commands from console
[root@localhost libexec]# ./check_rrdtraf -f /var/lib/mrtg/192.168.0.4_21.rrd -w 50,50 -c 80,80 -l M
OK - Current BW in: .03Mbps Out: .29Mbps|in=.036902Mb/s;50;80 out=.294127Mb/s;50;80
[root@localhost libexec]# ./check_ifoperstatus -H 192.168.0.4 -C public -k 21
OK: Interface C5 (index 21) is up.[root@localhost libexec]#

How they show up in advances status view

Advanced Status Details
Service State: Ok
Duration: 23d 23h 3m 56s
State Type: Hard
Current Check: 1 of 5
Last Check: 2010-03-30 19:18:39
Next Check: 2010-03-30 19:21:39
Last State Change: 2010-03-06 19:16:09
Last Notification: Never
Check Type: Active
Check Latency: 0.162 seconds
Execution Time: 0.03719 seconds
State Change: 0%
Performance Data: in=.036902Mb/s;50;80 out=.294127Mb/s;50;80

Which appears fine but again looking at the Performance graph for this


And this is a 10/100 Port on a HP switch.

Is there anything specific I can type into the console of the server to give you more information on this that might help me out here?

[mmestnik]

Ok,
The graphs are generated directly from the rrd files and not the output of check_rrdtraf. I'm not sure how much control over the process of generating graphs we have and the units of the data you are talking about. After looking over the PNP documentation I would guess that this could only be changed by modifying the source to PNP.

Perhaps this question should be asked on the PNP mailing list, they would know the why and hows for this question.
https://lists.sourceforge.net/lists/lis ... gios-users