[rrd-users] Re: Advanced CDEF Question

Dayton Turner dayton at digidefense.com
Fri May 28 19:01:41 MEST 2004

> On Wed, May 26, 2004 at 02:03:47PM -0600, Dayton Turner wrote:
>>   Anyways, Ive got this working, using the CDEF
>>   a,b,+,c,+,d,+,1024,*,8,* (to change to bits and mult by 1024)
>>   and
>>   e,f,+,g,+,h,+,1024,*,8,8
> that 2nd eight is a typo.  Yes, this looks good, except then NaN
> is involved.

Pardon me, the 8,8 WAS a typo, it was intended to read 8,*

>>   b,a,LT,0,b,a,-,IF
>> (e,f,+,g,+,h,+,1024,*,8,*),(a,b,+,c,+,d,+,1024,*,8,*),LT,0,(e,f,+,g,+, 
>> h,
>> +,1024,*,8,*),(a,b,+,c,+,d,+,1024,*,8,*),-,IF
> Are you putting those parentheses in for real?  You shouldn't.
> For readability I'd define two CDEFs and use those in the 3rd:
> CDEF:tmp1=a,b,+,c,+,d,+,1024,*,8,*
> CDEF:tmp2=e,f,+,g,+,h,+,1024,*,8,*
> CDEF:save=tmp2,tmp1,LT,0,tmp2,tmp1,-,IF

Yes I was putting those parentheses in for real. Can you tell me why  
that is not supposed to work? I ask because, as you point out later, i  
am infact using a frontend to manage my RRD files (Cacti). I took the  
old simpler CDEF of b,a,LT,0,b,a,-,IF and substituted in their  
interface "Use another CDEF" for where b was, and where a was,  
respectively. It created a CDEF using the parentheses. If another  
method is better, I'd like to pass that along to the authors and/or  
submit a patch myself. I see your use of variables in the above  
example, and currently as far as i know, Cacti will only allow me to  
use one CDEF per graph item, which is why, I presume, it tried to  
create a big long CDEF out of parentheses instead of split it across 3  
separate CDEFs.

My question aside from the above is, if I dont have an easy way  
currently to use 3 CDEFs in one graph item, can i somehow use this  
method of variable storing, or a different method of what i was  
intending to do with the parentheses, on just one CDEF?

>>   When i look at the graph in graph debug mode, It just shows a broken
>> graph, and "RRDtool Says: "
>>   Normally it would say OK, or if there was an error, spit out what  
>> the
>> error was. This is obviously making it hard to debug.
> That is a front end not doing its job I guess.

Yep, I've already submitted it as a bug to the Cacti developers. More  
info on why it might have done that, or how i could get the raw output  
from rrdtool (maybe the Cacti error parser ignored it) would be great,  
but I understand that you probably arent familiar with Cacti, so I'll  
bug them about that. :)


Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi

More information about the rrd-users mailing list