new installation, graphs showing NaN :(

[masdtaylor]

I just installed Cacti 0.8.8b on a new SLES 11SP3 install
and I am having issues getting any graph to show data.

I am trying to graph the throughput of certain interfaces on
our core switch (an HP 5820).

Instead of showing any data, my graphs just show nan.

I did some searching on google and this forum and came up with the
suggestion to follow this debugging guide.

http://docs.cacti.net/manual:087:4_help ... #debugging

I will outline the debugging troubleshooting I have performed and the results below

CHECK CACTI LOG FILE -
There are no instances of the "WARNING: SNMP timeout detected" in the log file

CHECK BASIC DATA GATHERING -
This device is gathering basic SNMP data, and when I run

"snmpget -c %ourpublic% -v 1 10.10.1.1 .1.3.6.1.2.1.2.2.1.16.56"

it returns

IF-MIB::ifOutOctets.56 = Counter32: %anumber%

should I be using our public or private password after the -c? both seem to work.

CHECK CACTI'S POLLER -
I cleared the cacti log (/var/log/cacti/cacti.log) and set the Poller logging level
to debug. I then found the host ID associated with the 10.10.1.1 switch. It was 4.
I then ran

php -q /usr/share/cacti/cmd.php 4 4

it gave no output to the screen, but when I checked the cacti.log, I found this

Code: Select all

02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] SNMP: Host responded to SNMP
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] RECACHE: Processing 2 items in the auto reindex cache for '10.10.1.1'.
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] RECACHE DQ[1] OID: .1.3.6.1.2.1.1.3.0
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] RECACHE DQ[1] OID: .1.3.6.1.2.1.1.3.0, output: 1505002104
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] RECACHE DQ[10] OID: .1.3.6.1.2.1.1.3.0
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] RECACHE DQ[10] OID: .1.3.6.1.2.1.1.3.0, output: 1505002105
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] DS[23] SNMP: v1: 10.10.1.1, dsname: traffic_in, oid: .1.3.6.1.2.1.2.2.1.10.56, output: 2117281014
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] DS[23] SNMP: v1: 10.10.1.1, dsname: traffic_out, oid: .1.3.6.1.2.1.2.2.1.16.56, output: 71789474
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] DS[24] SNMP: v1: 10.10.1.1, dsname: traffic_in, oid: .1.3.6.1.2.1.2.2.1.10.57, output: 135338944
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Host[4] DS[24] SNMP: v1: 10.10.1.1, dsname: traffic_out, oid: .1.3.6.1.2.1.2.2.1.16.57, output: 852177648
02/07/2014 07:39:44 PM - CMDPHP: Poller[0] Time: 0.0420 s, Theads: N/A, Hosts: 1

CHECK BULKWALK BEHAVIOUR (SNMP DATA QUERIES ONLY) -
I am already using SNMP v1, so BULKWALK doesn't apply, but just to be safe I changed my max OIDs to 1

CHECK MYSQL UPDATE -
I am not certain what debug log I should be checking for the failed MySQL update statement for the host concerning
table poller_output. I didn't complete this step. Any guidance on this one would be greatly appreciated.

CHECK RRD FILE UPDATE -
This one is the same as the SQL update, I'm not sure what debug log I should be looking in. I can tell you that
I found the RRD files located in

/usr/share/cacti/rra

I checked the timestamps on the files, and they don't change (nor does the size change) when I run the

php -q /usr/share/cacti/cmd.php 4 4

could this be what my problem is?

CHECK RRD FILE OWNERSHIP -
All the RRD files are owned by wwwrun:www, which I believe is what the poller is running as.
how would I check to see what the poller is running as?


This is happening for any graph that I make for my switches. Any help you can provide would be greatly appreciated.
Thanks in advance,

Taylor

[Shmegles]

A few quick things to check:
1. Is the poller enabled? Settings-->Poller and make sure "Enabled" box is checked.
2. Is each device enabled? Devices-->select a device and make sure the "Disable Host" box is not checked.
3. When using command line snmpget what user is running the command. Make sure you're wwwrun:www (or whomever the poller is using) to make that user can run 'snmpget'. Can also sudo to the correct user and type "which snmpget" and post the result.

[masdtaylor]

1) The poller is enabled
2) All devices are enabled
3) I have tested the snmpget using both root and wwwrun, both work great.
also, "which snmpget" returns the same thing if run by root or wwwrun
it's "/usr/bin/snmpget"

Update -
I made three test graphs last Friday. When I left Friday afternoon, all three were showing "-nan" instead of actual values. These graphs were made early in the day on Friday and had been running for hours.
When I came in today, two of those graphs had data for incoming and outgoing, and the third graph only had data for the outbound, but still showed "-nan" for incoming.

I'm not sure why these graphs aren't being populated almost instantly. Our poller is running every 5 minutes. I figured within 10 minutes of creating a graph it would be displaying data...

I have just continued to make our graphs, hoping that they will be populated with data tomorrow morning.

If anyone can give me some suggestions or guidance, I would appreciate it a lot

Thanks!

Taylor