No graph data for certain devices?

[jtzako]

I have a cacti server which was configured years ago by a previous admin. It has been working great on the devices we had been using all this time. However, we recently started using a new device, which is not working properly in Cacti.

Its a similar device (same mfg too) but the interfaces and CPU data etc are not graphing. Cacti poller did detect the device and added graphs for the CPU and various interfaces, but there is no data appearing on the graphs.

For example this one: "1/1/c1/1 Egress QoS - 1/1/c1/1, 100-Gig Ethernet, LAG-165 100" which does have quite a bit of traffic on it has nothing on the graph.

Im wondering if there is a difference in the SNMP that needs to be accommodated (different oid?), but I'm not sure where exactly that is being defined.

[macan]

enable device debug for problematic device and show log.

[jtzako]

enable device debug for problematic device and show log.

I went looking for logs and found this sort of thing on these new device:

2024-02-05 14:05:18 - SPINE: Poller[Main Poller] Device[OKLDCA033TW292948071LIO06] TH[1] DS[OKLDCA033TW292948071LIO06 - 1/1/c1/1 Egress QoS - 1/1/c1/1, 100-Gig Ethernet, LAG-163 100 ] Graphs[OKLDCA033TW292948071LIO06 - 1/1/c1/1 Egress QoS - LAG-163 100 GigE to SFFCA302SW8 PO923 11] WARNING: SNMP timeout detected [500 ms], ignoring host '<hostipomitted>'

The interesting thing is, cacti did poll these devices to get all the interface names listed, but appears to timeout when attempting to get data for them. Might that indicate the OID# is wrong? If so, where would I go to find where those are defined (we have thousands of devices, and there are templates, but not sure how that works).

[macan]

Type of device?
Which template is used for this device?

[jtzako]

Type of device?
Which template is used for this device?

Its a Nokia SR-1, the other devices are Nokia 7450 and 7750. There are some differences in how SR-1 works in some areas.
The template is a generic one I believe was created by the original admin. Its not a built-in cacti template.

[macan]

Ok, go to Utilities -> system utilities -> view poller cache -> select any working device. You will see used snmp OID.

You can try snmp get/walk this OIDs for not working devices. If nothinr or error is returned, so the device doesn't support this oid

[jtzako]

Ok, go to Utilities -> system utilities -> view poller cache -> select any working device. You will see used snmp OID.

You can try snmp get/walk this OIDs for not working devices. If nothinr or error is returned, so the device doesn't support this oid

I went into the poller cache and got the OID it uses for working and the non-working devices, and checked them on command line snmp walk and they work fine.

working device:
# snmpwalk -v 2c -c <community> <ipomitted> .1.3.6.1.4.1.6527.3.1.2.2.12.2.1.2.1.35749888.2
SNMPv2-SMI::enterprises.6527.3.1.2.2.12.2.1.2.1.35749888.2 = Counter64: 323532

non-working device:
# snmpwalk -v 2c -c <community> <ipomitted> .1.3.6.1.4.1.6527.3.1.2.2.12.2.1.4.1.1610899521.2
SNMPv2-SMI::enterprises.6527.3.1.2.2.12.2.1.4.1.1610899521.2 = Counter64: 2988621303

[macan]

Go to Templates -> Data sources -> choose template which is problematic -> in data source item check maximum value. Is there 'U' or sufficiently large value? It must be greater than 2988621303.

You can try tu run Troubleshooting -> Data sources too

[jtzako]

Go to Templates -> Data sources -> choose template which is problematic -> in data source item check maximum value. Is there 'U' or sufficiently large value? It must be greater than 2988621303.

You can try tu run Troubleshooting -> Data sources too

The max value for "U" is 4294967296 on inProfFwd_01-08 and OutProfFwd_01-08 in that template
Ill take a look at the Troubleshooting tab as well

[jtzako]

You can try tu run Troubleshooting -> Data sources too

Im getting 'data source returned bad results for' on all of the inproffwd and outproffwd entries.

[jtzako]

Go to Templates -> Data sources -> choose template which is problematic -> in data source item check maximum value. Is there 'U' or sufficiently large value? It must be greater than 2988621303.

You can try tu run Troubleshooting -> Data sources too

I did some digging around and found a setting for max SNMP OIDs per get request and changed it. The system was set to default to 60, but I noticed that the devices that were working had been set lower. I tried manually changing one of the non working devices down to 10 and now it is graphing.

[Osiris]

Increase timeout and reduce max oids. That'll get it for most devices.