Trying to do some server monitoring and failing miserably for anything other than standard stuff.
How do you use functions and ‘oids’, I have tried all kinds of things, help documentation isn’t very clear on the subject, actually it doesn’t cover it at all.
If I want to monitor CPU and Disk on a server for instance just how is it achieved, there is no ‘Disk or CPU’ service to discover - least ways none get discovered, these are standard windows boxes.
Then how do you consume those values / use functions etc, oids - how is one supposed to learn these things ???
I feel like an absolute thicko because clearly some here know how.
PS : SNMP is enabled and active on the boxes
What OS are you trying to monitor?
Windows: Make sure the IP address of the Dude server is listed as allowed to talk to the SNMP service
Windows: Windows firewall allowing SNMP traffic
Linux: SNMP is running and not locked down to local traffic requests?
Etc
Server 2008 R2, SNMP configured, set up a public community as read only, made sure it is permitted through the firewall, snmpwalk sees nothing … very frustrating.
I tried with some specific oids that I expect on server 2008 such as 1.3.6.1.2.1.25.2.3.1.6.6 (used memory) but still nothing, I set up a trap to catch authentication requests but I’m not seeing those either. I have verified that the firewall isn’t blocking anything on either port 161 or 162.
Ah 2008… ok… i know I have a few 2008 boxes that are what I’d call “odd”.
Sometimes they don’t show up at all… very frustrating.
Does it work without the firewall up? When I’ve had problems, dropping the firewall works fine. (not a solution but shows me it is the cause of my problems)
I have all the exceptions set up too - I have seen murmors of there being a few issues. (can’t find the links at the moment)
Don’t worry, doesn’t seem like you’re being thick 
Firewall up / down, domain, private or public settings make zero difference … just resorting to wireshark …
First I believe you would have had much more understanding if your w2k8 server played nicely and that is the trouble. Try adding some other devices… Take your time the dude is a great tool.
If you have not, you should read the probe thread to find probes you like. SNMPwalk is a great tool. Gsandul has some of the best posts around here as well. Once you determine your SNMP issue you should have little trouble.
So here is one probe broken down to bits to hopefully help you understand the steps to take to start building your own probes.
- If you are not certain which OID you are interested in Search Google. In this case I searched CPU CISCO OID.
- Once a rough idea of the OID is found use SNMPwalk on the device and verify that you can read the oids.
a. The oid for Cisco CPU and some other metrics is 1.3.6.1.4.1.9.2.1, use most of the oid you are interested in, not the entire oid. You will want to see all the values near the oid you are looking for.
- Once you have the exact oid place it in a new function… Give the function a name and place your custom if statement in the code.
if(array_size(oid_column(“1.3.6.1.4.1.9.2.1.57”, 10 ,5)), oid(“1.3.6.1.4.1.9.2.1.57.0”, 10, 5)+1 ,“False”)
Notice the array_size oid is not the entire oid but the returned value is the exact oid +1.
- Now take that new function and use it in a new probe.
a. Name the probe, Cisco CPU.
b. type of probe, Function.
c. Agent, default (unless you want a different dude server to execute the function).
d. Available, Cisco_CPU_a()<>False - Use the name of your function to call it and if available “the probe can be installed”.
e. Error, if(Cisco_CPU_a()<>“False”,if(Cisco_CPU_a() -1< 70, “”, concatenate("Warning: high CPU = ", Cisco_CPU_a(), “%”)), “Cisco Device down”) - This is complex but it reads "If the CPU can be read and if the CPU minus 1 is less than 70 return “” else return warning high CPU, else return device down - Notice that it is an if inside an if. We want to know if the device is down or if the CPU is too high and those will be placed in the notification email when done like this.
f. Value, Cisco_CPU_a() - graph this value.
g. Unit, % - This determines the graph units, every probe with the same unit are graphed on the same graph.
I hope that helps you out and gets you started on the right foot.
Lebowski
Regarding the Win2008 SNMP config. Have you configured the SNMP service after you installed it? Add a Read-Only community (public) and permit access to it from your Dude server (and whatever other devices you want to be able to query it) within the service configuration. You probably need to restart the SNMP service once you have made these changes as well (good old Microsoft). I have set up heaps of Win2K3 boxes and we have just started migrating to 2008 and have set up a couple of those so far. I have not had the issues you are faced with at this stage.
It seems to be as one poster has indicated, although the permissions are supposedly correctly set on the useless Windows firewall they don’t work, disabling the firewall service also kills the connection, the firewall service has to be running and the ‘domain’ firewall option turned off … ironically we also had to restart the service too after some changes.
Trying to find out why this is.
Trying to find a way to handle SNMP traps at the moment (parse trap and respond based on contents).