Any solution to negative RSSI returned from generic oid?

[ptera]

I have been trying to get rssi information from our Trango Access Points to graph the rssi levels of the customers radio. I have tried the templates listed elsewhere in this forum but they do not work. When using mrtg one can add 100 to the value being returned to make the number non-negative. I see no way to that here. I have tried setting min value to -100 but I still get "nan" on my graphs. I have looked at other listings here but I can find nothing helping. I want to make sure I have not missed anything before posting a bug report.

[BSOD2600]

Most likely it's because the data template you're using has the minimum value set to 0. You need to change this to -100 or something smaller. Then either delete the rrd file so cacti will recreate it or use rrdtool tune to fix the rrd file(s).

[ptera]

Man creating graph templates and such from scratch.....
A challenge it was but I think I just about got it licked...
Now I just have to repeat the same process about 400 times

Oh what fun

[gandalf]

Man creating graph templates and such from scratch.....
A challenge it was but I think I just about got it licked...
Now I just have to repeat the same process about 400 times

Oh what fun

That's not a deal with Graph Templates or Graphs but with a Data Template (if templated, should be only one). The rrdtool tune should not be that complicated, even with some 100's of rrd files. Or am I missing sth?
Reinhard

[ptera]

Here is the deal....I have to get my RSSI data using snmp oid. Simple except the data for each SU is polled from the Access Point and not from the SUs. It takes 2 oids, one for RSI level at the AP and one for the RSSI level at the SU. That is per SU (or per customer). Now we also want to put the two on their own graph. So we end up with a graph for each SU each with two oid. Mind you these all come from the same device or host. I had to modify the Generic SNMP to get two oids and then had to modfy a graph tempplate to show the two results of the two oids. And then I had to create two data sources to go with those. Man I am confusing myself. Anyway having the ability to duplicate some of these makes it easier since all I have to do is change oid to match which SU I am trying to graph. Did that make any sence?

[gandalf]

I would not object that this approach works or is best for you.
Nevertheless, I'd suggest discussing another approach:
- assume you create some wrapper script (e.g. in perl, php, even sh if you want to) taking the specific OIDs as parameters, one for RSI level at the AP and one for the RSSI level at the SU
- this script should spit out the snmpget-ed data e.g. as rsi:... rssi:...
- then, a single Data Template defining two data sources, rsi and rssi would hold the data
- last, a single Graph Template, defining at least two graph items (8 if you prefer using whole legend GPRINTs) would do the trick
You may have a read at the second to last link of my signature for a walkthrough including screenshots.
If I missed sth, feel free to make it clear
Reinhard