Hello,

I am running Cacti 1.1.38, most of my graphs are graphing correctly. I do have one set that are not graphing any data. I did notice that the .rrd files are not updating for those graphs, I have gone through the debug steps but no luck.

I can do a snmpwalk and snmpget to the the server with the oid. I did try to manually update an rrd but it did it once and will not continue to do so, I continue to get -nan when I fetch the .rrd files from that server. I have the correct permissions for the graphs.

Any help would be great...Thansk

when I run the fetch I get.
1546014000: -nan -nan
1546014060: -nan -nan
1546014120: -nan -nan

Permissions + poller cache inspection. Last, put the device into debug mode and look for errors.

I think I found my problem, I am currently polling two servers that are a like. When I rebuild the device for one of them it starts working and the other one stops, and vice versa. I can only get one of the two servers to run at one time, all of my other servers are graphing as expected.

I tried to change the name and the community string on the devices but once one of them are working the other one quits.(strange) I can manually update the rrd file but then It will not continue to poll it.

Any ideas???

Check the hostname, if the hostname is different, check that they are not resolving to the same machine (ie, if the IP is the same or if the same machine is hosting both IPs).

NetniV,

They are two different servers, and both have a different IP, they are on the same subnet, have the same naming structure, and have the same community string. That's when I changed the naming and the community string to see if it would help but it did not.

I have an old graphing server that is graphing both of those servers and I merely copied the devices on mine to match it exactly but no luck. This old server is running cacti 0.8.8b, and my new one is running cacti 1.1.38.

I am pulling with cmd.php, do you think switching to spine would make a difference?

Also when I put it into debug mode and watch the logs it only shows the one that it is pulling the one that is having issues dose not show up in the log unless I manually pull it. Then it will show up one time and then done.

thanks

If you look at the device status, do they both show as up? What device status detection method are you using?

Yes they both show as up, and shows the correct up-time. I am using SNMP - Interface Statistics with an re-index method of up-time. Both give me an success status.

SNMP details are held against the device. Your data sources can only be associated with one host. Therefore, I would suggest debugging one host at a given time.

Edit one of the devices, verify the SNMP details (make sure to save any changes) then enable the debug option at the top and then check the log files for the responses. When you are happy that device is working, disable the debug and move onto the other device, rinse and repeat.

NetniV,

I ended up installing spine, and this solved my issue. I would put the device in debug but it didn't give any errors or problems with that device, if I left them both on and put one in debug it wouldn't give me any info on it at all. I was a strange problem but with spine everything is working fine.

Thanks for your help!