status Unknown with SNMP Downed Host Detection

fitz · Post by **fitz** » Wed Mar 21, 2007 4:59 pm

Hello,

I have been using SNMP for Downed Host Detection for the last year and it has been working flawlessly. Recent I tried to added several new hosts and cacti is able to poll the host, but the Status says Unknown even though the Availability says 100%.

It's seems that when the status us Unknown no graphs are created or updated for that host which makes sense.

I have tried manually snmpwalking the hosts in question from the cacti server and get back results so I know it's not a firewall issue.

Can anyone help shed some light on this?

Thanks,
Sean

maw · Post by **maw** » Wed Mar 21, 2007 5:56 pm

Hi Fitz,

When you goto:
devices > device showing unknown >

does the new window show SNMP error in the top left?

fitz · Post by **fitz** » Wed Mar 21, 2007 6:39 pm

Hi,

I do not see any SNMP errors for the host. It is able to get the basic info and it finds the usual items to graph (disk, network,etc).

I have attached a couple screen captures.

-Sean

fmangeant · Post by **fmangeant** » Thu Mar 22, 2007 2:40 am

Hi

are you sure that the poller is running ?

Can you run it once by hand ? The status should then be "recovering" or "down".

fitz · Post by **fitz** » Fri Mar 23, 2007 12:45 pm

The poller is running as other graphs are updating and the cacti.log shows activity.

I can run it by hand, but it still shows this host a being unknown.

My install is on RHEL 4 with cacti-0.8.6h, cacti-cactid-0.8.6i, cacti-plugin-arch and net-snmp-5.1.2-11.

It has been working great until recently. No changes have been made except for trying to add new hosts.

-Sean

kdawg44 · Post by **kdawg44** » Mon Mar 26, 2007 1:48 pm

I am having the same issue except mine only occurs with Windows Hosts. My firewall and Linux hosts are up but the windows hosts show unknown even though they are walkable and are getting the SNMP information.

THanks for any suggestions.

Post by **gandalf** » Thu Mar 29, 2007 2:08 pm

"unknown" is the initial status of a new device. It will not change until
- at least a single graph is defined for this host
- the poller sees this host as up (depending on "downed host detection")
Reinhard

fitz · Post by **fitz** » Tue Apr 03, 2007 2:20 pm

Is it possible that the poller can't complete polling before cron kicks off another job?

I now understand what changes Status, but I can't understand why the systems are stay unknown. As the availbilty is 100% and I can manually snmpwalk the server and get the results, but no graphs are created.

-Sean

Post by **gandalf** » Tue Apr 03, 2007 2:39 pm

Please follow hints of second link of my signature. This is some generic debugging stuff. I hope it will help you
Reinhard

Post by **TheWitness** » Thu Apr 05, 2007 4:33 am

Run your poller in LOW and post a few log entries. Unknown will continue if you are using "cmd.php". If you use "cactid", then the host will appear Up/Down immediately, even if you don't have graphs.

If a device is not being polled due to timing, your overall polling will be right around 300 seconds due to the poller being terminated. That will show in your posted log.

This minor difference should be noted. Otherwise, Gandolf is correct.

TheWitness

fitz · Post by **fitz** » Thu Apr 05, 2007 4:29 pm

I found switching from cactid to cmd.php fixed the problem.

I'd like to get working with cactid again.

I'll port some logs.

-Sean

Post by **TheWitness** » Thu Apr 05, 2007 8:17 pm

Cool. The ball's in your court.

TheWitness

bluephile · Post by **bluephile** » Mon Apr 09, 2007 2:53 am

I have the exact same problem as fitz. I'm using cmd.php. The logs show simply:

Code: Select all

SYSTEM STATS: Time:293.2785 Method:cmd.php Processes:2 Threads:N/A Hosts:5 HostsPerProcess:3 DataSources:17 RRDsProcessed:5
POLLER: Poller[0] Maximum runtime of 292 seconds exceeded. Exiting.
SYSTEM STATS: Time:293.2683 Method:cmd.php Processes:2 Threads:N/A Hosts:5 HostsPerProcess:3 DataSources:17 RRDsProcessed:5

etc. etc.

Other hosts are working fine. I'm using SNMP & UDP Ping host downed detection. The appropriate SNMP info is showing on the device details page. Several graphs have been added for this host, but the rrds aren't being created.

Thanks,
bluephile

EDIT: I should mention that manually running the commands shown on the data source pages works; the rrds are created and the graphs are displayed. The rrds are only updated when I run the commands, however.

EDIT 2: I should also mention that all the other hosts that I'm monitoring are Linux & Windows, but this is an AIX host running net-snmp. Thanks!

Post by **TheWitness** » Mon Apr 09, 2007 7:05 am

Please insure you are running all the latest patches and that your path to snmpgetnext is valid.

TheWitness

bluephile · Post by **bluephile** » Wed Apr 18, 2007 7:25 pm

Just a status update, in case anyone stumbles across this: switching to SNMP version 2 for the host magically fixed the problem.

Thanks,
bluephile

Cacti

status Unknown with SNMP Downed Host Detection

status Unknown with SNMP Downed Host Detection

Who is online