dhcpd-snmp, too many graphs?

[SaracenMan]

I started using dhcpd-snmp to monitor my new DHCP servers as I migrated subnets over from the old ones. As each subnet was migrated, I added it to cacti and all was happy.

Around 50 odd subnets and the new ones no longer appeared.

I now have the existing graphs continuing to work, but the device display shows "Sucess [0 Items, 0 Rows]".

Using snmpwalk on the cacti hosts shows that there is no problem fetching all the data. Its all there.

Verbose Query shows:

Running data query [11].
+ Found type = '3' [snmp query].
+ Found data query XML file at '/cacti/resource/snmp_queries/dhcpd-snmp.xml'
+ XML file parsed ok.
+ Executing SNMP walk for list of indexes @ '.1.3.6.1.4.1.21695.1.2.2.1'
+ Located input field 'dhcpdSnmpPoolIndex' [walk]
+ Executing SNMP walk for data @ '.1.3.6.1.4.1.21695.1.2.2.1'
+ Located input field 'dhcpdSnmpPoolDescription' [walk]
+ Executing SNMP walk for data @ '.1.3.6.1.4.1.21695.1.2.2.2'
+ Found data query XML file at '/cacti/resource/snmp_queries/dhcpd-snmp.xml'
+ ... Repeated 186 times.

Walking those OIDS results in the correct data too.

Cacti Version: 0.8.7a

Any ideas?

[gandalf]

Please see 2nd link of my sig for debugging
Reinhard

[SaracenMan]

Thanks for the debugging page.

Tried all that and there are no obvious errors.

It all seems to be working, but not for all the items the server has to offer..

There are too few rrd files in the rra directory for the number of DHCP pools the servers have and report via SNMP:

% snmpget -c public -v 2c server1 .1.3.6.1.4.1.21695.1.2.1
SNMPv2-SMI::enterprises.21695.1.2.1 = INTEGER: 55

% ls server1_dhcpdsnmppoolavaila* | wc -l
46
% ls server1_dhcpdsnmppoolactive* | wc -l
46
% ls server1_dhcpdsnmppoolexp* | wc -l
46
% ls server1_dhcpdsnmppoolsi* | wc -l
46

11 Files Missing!

% snmpget -c public -v 2c server2 .1.3.6.1.4.1.21695.1.2.1
SNMPv2-SMI::enterprises.21695.1.2.1 = INTEGER: 51

% ls server2_dhcpdsnmppoolavaila* | wc -l
22
% ls server2_dhcpdsnmppoolac* | wc -l
22
% ls server2_dhcpdsnmppoolex* | wc -l
22
% ls server2_dhcpdsnmppoolsi* | wc -l
22

29 Files Missing!!!


Something somewhere stopped correctly adding in new items. If both servers had stopped at the same value, it would have looked like a limit had been reached, but 46/55 and 22/51 does not indicate this?

Any clues.

Thanks.

[gymmarn]

Try changing the timeout in Cacti SNMP Timeout to some high amount (I changed from 500 to 30000).

You can check this phenomenon if you try snmpwalking from the Cacti machine to the DHCP-server. If you get all the items slowly it means that it's all working you just need to get Cacti to wait abit for the server to hand over the information

For example (might give timeout):
snmpwalk IP -c PASS .1.3.6.1.4.1.21695.1.2 -v2c

Wait 30s might work, if so then use this with Cacti aswell.
snmpwalk IP -t 30 -c PASS .1.3.6.1.4.1.21695.1.2 -v2c

Hope this helps!

[SaracenMan]

Yep. I found this. It was the SNMP timeout value. Increased and problem went away!

Thanks for replying.