Data Souce Custom Data Section, is this correct?

[imthemp3king]

I have created a data query, data template, and graph template to be used to graph the temperature probe readings in my Dell servers. When I turn on Data Source debugging, I get this output which seems to be like most of my other data sources:
/usr/bin/rrdtool create \
/var/www/cacti/rra/xxxx_xx_xx_xxx_mobo_temp_105.rrd \
--step 300 \
DS:mobo_temp:GAUGE:600:0:1000 \
DS:cpu1_temp:GAUGE:600:0:1000 \
DS:cpu2_temp:GAUGE:600:0:1000 \
DS:bckpln_hrddsk_temp:GAUGE:600:0:1000 \
DS:bkpln_ambt_temp:GAUGE:600:0:1000 \
RRA:AVERAGE:0.5:1:600 \
RRA:AVERAGE:0.5:6:700 \
RRA:AVERAGE:0.5:24:775 \
RRA:AVERAGE:0.5:288:797 \
RRA:MAX:0.5:1:600 \
RRA:MAX:0.5:6:700 \
RRA:MAX:0.5:24:775 \
RRA:MAX:0.5:288:797 \

When I go into Graph Management and turn on graph debugging to view the graph I get:
RRDTool Says:

ERROR: opening '/var/www/cacti/rra/xxxx_xx_xx_xxx_mobo_temp_105.rrd': No such file or directory

I have checked the permissions on the rra and log directories and cacti has rw permissions. The only thing that really stands out to me is this part of the graph management screen for the particular device I am trying to monitor temperatures on

Is this correct? This seems to be the only data source that has Data query data sources must be created through New Graphs. next to Index Type and Output Type ID

[melchandra]

What kind of template is your data source using? Are you just doing static oids. Because then you don't need the dynamic fields (the 3 fields at the bottom). Just remove them from the Data template. Those fields are for doing traffic interfaces where you give it the root oid of the interface you want to monitor - and these 3 fields control how it is indexed, and which paticular index you are monitoring. Looking at what you are trying to graph, I'm pretty sure this is not what you want. Look at the Data template for Linux CPU usage I think that is closer to what you want.

[imthemp3king]

Here are the templates and the XML file I used

[imthemp3king]

Here is the XML file

[imthemp3king]

Then should I unchek the same 3 boxes in the Data Query?

[melchandra]

no, I think you need to just change the data input method from "Get snmp data (indexed)" to just plain "Get snmp data".

[imthemp3king]

I did that and now I get the graph but the results are all Nan. I ran cmd.php manually and this was the result:
snmp: xxxx.xx.xx.xxx:161, dsname: bckpln_hrddsk_temp, oid: , value:update /var/www/cacti/rra/xxxx.xx.xx.xxx_mobo_temp_117.rrd --template bckpln_hrddsk_temp N:U
I have verified that the actually readings can be obtained by walking the tree
snmpwalk -v 1 -c xxxxxxx xxxx.xx.xx.xxx .1.3.6.1.4.1.674.10892.1.700.20.1.6
SNMPv2-SMI::enterprises.674.10892.1.700.20.1.6.1.1 = INTEGER: 220
SNMPv2-SMI::enterprises.674.10892.1.700.20.1.6.1.2 = INTEGER: 220
SNMPv2-SMI::enterprises.674.10892.1.700.20.1.6.1.3 = INTEGER: 230
SNMPv2-SMI::enterprises.674.10892.1.700.20.1.6.1.4 = INTEGER: 170
SNMPv2-SMI::enterprises.674.10892.1.700.20.1.6.1.5 = INTEGER: 220

[melchandra]

You probably have to go in and re-specify the oid you want monitored since you switched from an index mode to a non-indexed mode.

[imthemp3king]

where do I specify the OID? I have that OID specified as my output in the XML file. Are you referring to the data template (i just added the OID to it for the screenshot below, it isn't there at the moment)? If I'm not telling it to get non-indexed data and I am graphing 5 different data sources (each with their own OID) how does cacti correctly determine which data source gets the returned values? I apologize if I sound confused, I am used to creating simple .cfg files for MRTG and this is seems a little complicated to me. And this is the first time I have tried to create a graph with multiple data sources, so it makes it even more confusing for me. Thanks for the reply's

[melchandra]

Ok. Here is the deal. Some devices have multiple things all under the same branch of the oid tree. So let's say a device has 4 temperature monitors. Let's say the oid for temperatures is .1.3.6.1.999. Ok, so your first oid for the first temperature is .1.3.6.1.999.1, and the last is .1.3.6.1.999.4. There are basically 2 ways to approach this. We were kinda talking about differant ways and that I think is why we are both a little confused.

1. If you dell servers all have differant amounts of temperature monitors, or you want to expand this system to monitor differant temperatures, then you want to set it up the way you origionally had it, with the data query, and the indexed stuff. You want to copy the way the Traffic / Interface stuff is set up, from the Data Query, the data templates, and the graph templates. This is so that you can pick / choose which temperatures you want to monitor, so if one device has 7 temperature, another 3, another 4, then this is what you want. You run the query on .1.3.6.1.999, and Cacti will find everything directly under that oid. So .1.3.6.1.999.1, to .1.3.6.1.999.X (where X is the number of temperature sensors that the host has). What you also need is another oid to index them off of. Or the name them. The traffic / interface uses the ifDesc, which is generally another oid branch, let's say (for my little example) it's .1.3.6.1.998. So for each 999.X entry, there is a 998.X entry matching it with the associated description. So you need to figure out what two oid branches you are going to use for this.

2. The second way is static. If every server has a 3 temperature sensors at .1.3.6.1.999.1 .1.3.6.1.999.2 and .1.3.6.1.999.3, and you want to monitor all 3 sensors on every device. Then you want to set up a Graph with three data sources. Each data source will point directly to one of the OID's. This works because the hosts all have the same number of sensors, located at the same oid. This is the way I recommended because it sounded to me like this was more your situation.

I may have been wrong though. Hopefully this clears things up for you. I'm sorry if I made it more complicated, hopefully between you and the BB here, we can get things figured out for you. :-) Good Luck :-) Let me know how I can help more.

To answer your question on the post - the oid you want to change is right there on the bottom. You have tabs there for each one of your data sources. The annoying thing about this is you can only make changes to one data source, then you have to click save. If you click on another data source, the changes you've made to the current one are lost.

[imthemp3king]

Thanks so much for your assistance. It work fine now that I understand how to properly create the templates. I stopped trying to use the data query and just went with statically defined OID's and graph and data templates. Thanks again