I've hit the weirdest bump with cacti, while polling one of my servers. I've recently started out with a fresh install of cacti-1.1.31, which I've later upgraded to the latest 1.1.37 hoping it would fix this, but it seems the problem persists. I've populated the devices list with a few devices, and it seems all of them display the interface information (In/Out bits-64-bit counters and In/Out Unicast Packets) correctly, but one server doesn't:
I'm using cacti with spine, so the first step in troubleshooting was to enable debug logging and see what's going on. Related to this particular server, I'm seeing the following:
Code: Select all
2018-03-30 16:28:04 - SPINE: Poller[Main Poller] Device[5] Device[leviathan.<obfuscated>] TH[1] Graphs[leviathan.<obfuscated> - Traffic - eth0, leviathan.<obfuscated> - Traffic - eth0] DS[leviathan.<obfuscated> - Traffic - <public-ip-address> - eth0] SNMP: v2: <public-ip-address>, dsname: traffic_out, oid: .1.3.6.1.2.1.31.1.1.1.10.2, value: 529355196684
2018-03-30 16:28:04 - SPINE: Poller[Main Poller] Device[5] Device[leviathan.<obfuscated>] TH[1] Graphs[leviathan.<obfuscated> - Traffic - eth0, leviathan.<obfuscated> - Traffic - eth0] DS[leviathan.<obfuscated> - Traffic - <public-ip-address> - eth0] SNMP: v2: <public-ip-address>, dsname: traffic_in, oid: .1.3.6.1.2.1.31.1.1.1.6.2, value: 366238621169
The next step in troubleshooting, I turned on Graph Debug Mode, thinking there was something wrong with the rrd file. To my surprise, rrdtool reported everything is OK:
The destination server (leviathan) is a HP DL180 G6 server running ubuntu 16.04.4 LTS, and snmp 5.7.3+dfsg-1ubuntu4.1 from the repos. The network interfaces are polled correctly from the server, in the "create graphs for this device" page:
As a special mention, I've also created graphs for the dockerbr0, tun0, and tun12 interfaces, all of which pass traffic and neither display In/Out bits or In/Out Unicast Packets.
I've checked the cacti installed directory, everything is owned by the "www" user, so there's no permissions problems. I've also checked the cacti and webserver logs, don't have any warnings/errors. I'm polling a bunch of other servers, different models, physical and virtual and they all work. All except for this particular one.
I'm at a loss as to what to try next, can anyone throw a hint my way ?
P.S: As a special mention, other graphs (like processor, memory, disk usage, etc) work fine for this server. It seems only the "network interface"-related graphs don't work.
Thank you!
Regards,
-tbone