Switch monitoring wizard claims "no ports detected..."

[GldRush98]

Ok, back story;
All switches are 3com series, 4500 for this case.
I have ~20 other 4500 series 3coms added in, using the switch wizard to add them. They have all added in and given Nagios the info on their ports (port status and bandwidth).
One switch doesn't appear to be giving Nagios any info on its ports:

No ports were detected on the switch. Possible reasons for this include:

* The switch is currently down
* The switch does not exist at the address you specified
* SNMP support on the switch is disabled

So I run the command through a terminal session

Code: Select all

cfgmaker --show-op-down --zero-speed=100000000 --snmp-options=:::::1 --noreversedns [email protected]

And I get back what appears to be information on all of the switch ports (I have attached the output of the cfkmaker command to this post)

Something else I should mention; This switch was previously added in to Nagios with only a ping command working on it. There was an issue with SNMP not running on the right IP on the switch, but that was resolved on Friday. I have since verified the switch is properly replying to SNMP using a different plugin check command at the Nagios terminal:

Code: Select all

[root@localhost libexec]# ./check_3com_health.pl -H 10.20.1.28
3COM_HEALTH OK - FAN:OK  PWR:OK  MEM:OK  CPU:OK | fan=100%;100;76;0;100 pwr=100%;100;76;0;100 mem=38%;85;90;0;100 cpu=16%;85;90;0;100

When I removed the switch I went in to Core Config manager, deleted its services, then deleted the host, and applied the configuration (successfully).

So, I am not sure why Nagios is telling me there are no detected ports, when the manually ran command is showing me there are? Plus as I said, I have 20 other switches exactly like this added in with out issue. I'm stumped.
Any ideas?
Thanks,
Nick

[nscott]

I have had a similar problem on Cisco 29XX switches that were resolved with using a RW community string, are you currently using a RO?

[GldRush98]

The public community string is read-write. Just out of curiosity I switched it to a read-only community string, but that didn't make a difference.

What's more, if I use the SNMP walk wizard, it is showing a bunch of the interfaces...

snmpwalk.JPG

[nscott]

Can you try running this from the nagios/libexec/ ?

./check_snmp_int.pl -H 10.20.1.28 -C <insert_your_comm_string> -n

Does that return any interfaces?

[GldRush98]

Can you try running this from the nagios/libexec/ ?

./check_snmp_int.pl -H 10.20.1.28 -C <insert_your_comm_string> -n

Does that return any interfaces?

Yes, it does

Code: Select all

[root@localhost libexec]# ./check_snmp_int.pl -H 10.20.1.28 -C public -n
Ethernet1/0/16:DOWN, Ethernet1/0/24:UP, Ethernet1/0/18:UP, Ethernet1/0/4:UP, GigabitEthernet1/0/52:DOWN, Ethernet1/0/9:DOWN, Ethernet1/0/35:DOWN, Ethernet1/0/28:UP, Ethernet1/0/13:DOWN, Ethernet1/0/42:UP, Ethernet1/0/14:DOWN, Ethernet1/0/34:DOWN, Ethernet1/0/20:DOWN, Ethernet1/0/32:DOWN, Ethernet1/0/30:UP, GigabitEthernet1/0/50:DOWN, Ethernet1/0/12:DOWN, Ethernet1/0/47:DOWN, Ethernet1/0/2:DOWN, Ethernet1/0/38:DOWN, Vlan-interface5:UP, Ethernet1/0/15:DOWN, Ethernet1/0/43:DOWN, Ethernet1/0/44:UP, Ethernet1/0/39:DOWN, GigabitEthernet1/0/49:UP, Ethernet1/0/10:UP, Ethernet1/0/31:UP, Ethernet1/0/23:DOWN, Ethernet1/0/17:DOWN, Ethernet1/0/48:UP, Aux1/0/0:DOWN, Ethernet1/0/11:DOWN, Ethernet1/0/46:UP, Ethernet1/0/40:DOWN, Vlan-interface28:UP, GigabitEthernet1/0/51:DOWN, NULL0:UP, InLoopBack0:UP, Ethernet1/0/29:UP, Ethernet1/0/36:DOWN, Ethernet1/0/33:DOWN, Ethernet1/0/6:DOWN, Ethernet1/0/21:DOWN, Ethernet1/0/37:DOWN, Ethernet1/0/1:DOWN, Ethernet1/0/8:UP, Ethernet1/0/45:UP, Ethernet1/0/41:DOWN, Ethernet1/0/3:DOWN, Ethernet1/0/27:UP, Ethernet1/0/26:UP, Ethernet1/0/19:DOWN, Ethernet1/0/7:DOWN, Ethernet1/0/5:UP, Ethernet1/0/22:UP, Ethernet1/0/25:DOWN, Vlan-interface9:UP: 34 int NOK : CRITICAL
[root@localhost libexec]#

[GldRush98]

Hmmm, this is interesting...

Code: Select all

[root@localhost libexec]# locate 10.20.1.28
/etc/nagiosql/backup/hosts/10.20.1.28.cfg_old_20110613082727
/etc/nagiosql/backup/hosts/10.20.1.28.cfg_old_20110613104824
/etc/nagiosql/backup/hosts/10.20.1.28.cfg_old_20110613105335
/etc/nagiosql/backup/services/10.20.1.28.cfg_old_20110613082710
/etc/nagiosql/backup/services/10.20.1.28.cfg_old_20110613104903
/etc/nagiosql/backup/services/10.20.1.28.cfg_old_20110613105327
/usr/local/nagios/share/perfdata/10.20.1.28
/usr/local/nagios/share/perfdata/10.20.1.28/3com_health.rrd
/usr/local/nagios/share/perfdata/10.20.1.28/3com_health.xml
/usr/local/nagios/share/perfdata/10.20.1.28/Ping.rrd
/usr/local/nagios/share/perfdata/10.20.1.28/Ping.xml
/usr/local/nagios/share/perfdata/10.20.1.28/_HOST_.rrd
/usr/local/nagios/share/perfdata/10.20.1.28/_HOST_.xml
/usr/local/nagiosxi/tmp/mrtgscan-10.20.1.28
/usr/local/nagiosxi/tmp/mrtgscan-10.20.1.28.done
[root@localhost libexec]#

Even though I currently have 10.20.1.28 completely removed from Nagios via web interface...
It's clear some of those are old files and some might just be left over perfromance data files, but these two have me interested:
/usr/local/nagiosxi/tmp/mrtgscan-10.20.1.28
/usr/local/nagiosxi/tmp/mrtgscan-10.20.1.28.done

[GldRush98]

Further investigating:

Code: Select all

-rw-r--r-- 1 nagios nagios    561 May 11 13:42 mrtgscan-10.20.1.28
-rw-r--r-- 1 apache nagios      0 Jun 13 13:07 mrtgscan-10.20.1.28.done

The owner/datestamp were looking off to me, so I just renamed the files

Code: Select all

[root@localhost tmp]# mv mrtgscan-10.20.1.28 mrtgscan-10.20.1.28.killed
[root@localhost tmp]# mv mrtgscan-10.20.1.28.done mrtgscan-10.20.1.28.done.killed

And rescanned on the switch wizard, and BAM! It worked!! It pulled in all of my switch ports for status and bandwidth!

[nscott]

GldRush98,

Thats great! I'm glad you got it working, I will certainly keep this resolution in mind when troubleshooting in the future, thanks for sharing your troubleshooting steps.