Views-Service Details shows netsnmp error for all devices

[mfernacz]

If a device (switch) is unreachable, that status is reported for all switches. I used the network/router wizard to add these 2 devices. I then added 2 configuration files in NRDS for each device. If I cause one switch to be unreachable it is reported for both devices.

Am I doing something wrong?

[abrist]

I have to check - you did update the nrds config after making the IP changes, correct?

[mfernacz]

Thanks. I did not make any IP changes. I added the NRDS configuration, then install the client for both switches after I added them.

command[__HOST__]=/usr/local/nagios/libexec/check_ping -H 192.168.58.10 -w 200.0,40% -c 400.0,80% -p 1
command[Check Load]=/usr/local/nagios/libexec/check_snmp_load.pl -H 192.168.58.10 -C public -w 3,3,2 -c 4,4,3 -T netsl -f
command[Check Memory]=/usr/local/nagios/libexec/check_snmp_mem_mrv.pl -H 192.168.58.10 -C public -w 65,65 -c 90,90 -f

cd /tmp
wget -O 192.168.58.10.tar.gz "http://192.168.54.253/nrdp/?cmd=nrdsget ... .168.58.10"
gunzip -c 192.168.58.10.tar.gz | tar xf -
cd clients
./installnrds HOSTNAME INTERVAL

I then made 192.168.58.11 unreachable and got these results.

[abrist]

PM me your configuration snapshot tarball. To download the tarball, go to Admin --> Config Snapshots. Click the topmost "disk" icon, save the file, and send to me in a pm. Thanks.

[mfernacz]

Here is the latest.

[abrist]

Do you have any objects under --> Admin --> Unconfigured Objects?
I ask because the checks in question look like snmp traps, not NRDS checks. Did you setup snmp traps on this XI server?

[mfernacz]

I did not setup snmp traps. I assumed that if you cannot reach one device, it would not affect other service details for other devices.

[sreinhardt]

I noticed your post above stating that you made these hosts unreachable. Are they presently available for your nrds servers to check? From the nrds server try the following:

Code: Select all

nmap -p 161 192.168.58.10
nmap -p 161 192.168.58.11

[mfernacz]

If I pull the cable on Core2_11_Even which is 192.168.58.11, it shows it is not reachable which is what I expect. I do not expect the Core1_10_Odd (192.168.58.10) to show an error in Check Memory. If I run nmap -d command it works on 192.168.58.10, but fails on 192.168.58.11. The nmap command works on 192.168.58.11 when I put the cable back in. As I watch Check Load and Check Memory for both switches it has the same values. Not sure why because each host has their own NRDS configuration file.

[slansing]

It is of course totally possible that the trap reporting unreachable was able to be sent in time where the other server was not able to send out a trap about a change in it's check before it went completely down. This would be more of a system dependent cause than anything as the speed at which the check is occurring could have crossed between when it was brought off it's network connection.