SNMP Issue

[netniV]

When you create a device, you specify the SNMP settings against it. You can update these individually, or select a few devices and update in mass. You MUST specify the correct SNMP version, community and if required auth details. The remote server must allow SNMP port through its firewall and be configured to receive on the IP used by Cacti.

[MasonL]

Went through and checked everything. Port 161 is open through windows firewall for SNMP on all servers. All instances have the proper IP address of the server/servers they're connected to. I went through the Paessler tester and looked at my registry to make sure I am using the right snmp version aswell as community in my case. Clearly I have a config issue, I have been trying to go back through the documentation and double check my settings against it with no success. I may have messed up the receiving IP, where do I find the config for that?

[MasonL]

Went around to look for any other weird behavior, found some (hopefully) useful data. As seen below my cacti host CPU graph debugging log shows that the graph should be outputting some sort of data but is not.

CPUusageLOG.PNG (8.92 KiB) Viewed 1286 times

CPUusageGraph.PNG (40.79 KiB) Viewed 1286 times

Seems abnormal, no harm in pointing out seemingly odd behavior imo.

[camerabob]

From the run bar type 'services.msc' to open the services snap-in.

step1.png (55.45 KiB) Viewed 1282 times

From there go to SNMP Service and select properties

step2.png (62.21 KiB) Viewed 1282 times

After that cruise over to the 'Security' tab and fill in the blanks. I wiped out my information, but I'm sure you get it.

step3.png (47.24 KiB) Viewed 1282 times

Close all and test again. For localhost, use 127.0.0.1 for IPv4.

[MasonL]

Ok, this is something I have done already, glad to see I can read. Is there anything I need to do with the SNMP trap service below it?

[MasonL]

Went through my test servers, verified and double-checked everything was alright and replaced 'localhost' with '127.0.0.1' in each. After going through it did actually populate the graphs tab with all my graphs, went from 8 --> 48 graphs but very few actually contain data. The rest just show the '-nan(ind) %' value you can see on the CPU Utilization graph I attached previously. Below I am going to attach a current view of my logs where it shows the snmp returns as U values or "invalid response(s)"

073019.PNG (56.52 KiB) Viewed 1276 times

PS. If you see this tomorrow Happy Fourth!

[MasonL]

After reading just about every post made in the past year, I think the issue for me lies somewhere in the install. Every other setting I've gone from the ground up and made sure was correct, except for those involving the install. I already had mySQL 5.7 on the cactihost and couldn't delete to reinstall because it was already being used. When I first did the install my boss went through and "set up" the portion in mySQL which included making the schema and, well the issue there is I'm not sure what else was done and he doesn't remember exactly at this point. The only graphs/templates I am getting results on are my Poller Items, Poller Time, Main Poller DS/RRD, processes, logged on users, and uptime.

I ran through automation and cacti setup one of the servers I already have with a different IP, and the SNMP information in the device menu loads! The only issue is I still only get the graphs I mentioned above... That device has all the same SNMP community, version, port, device template setup identically to the other devices I have already. So after about a week of diving I'm still about as lost