[rrd-users] Using LAST on a CDEF

Chris Malton chrism at cmalton.me.uk
Mon Feb 25 00:18:49 CET 2013


Hi Yannick,

Will give that a go tomorrow.

Steve Shipway has come up with another solution off-list, and that seems 
to be approximately what I wanted - but you make a valid observation, 
and I will need to fix that.

Regards,

Chris

On 24/02/13 23:07, Yannick Marquet wrote:
> Hello,
>
> I think it's cause by this line :
> CDEF:ds4ak=ds4a,UN,0,ds4a,IF
> CDEF:ds4bk=ds4b,UN,0,ds4b,IF
> CDEF:totala=ds0ak,ds1ak,+,ds2ak,+,ds3ak,+,ds4ak,+
> Your CDEF make all UNK value are equal to 0, including uncollected 
> data (data collect in the future)
> Be careful, because if your display range ends after your last entry. 
> Your UNK (uncollected) values are replace by 0.
> So using LAST return 0.
>
> Try to use ADDNAN 
> (http://oss.oetiker.ch/rrdtool/doc/rrdgraph_rpn.en.html#IArithmetics) 
> instead of "+" and delete this line.
> CDEF:ds4ak=ds4a,UN,0,ds4a,IF
> CDEF:ds4bk=ds4b,UN,0,ds4b,IF
> Yannick Marquet.
>
>
>
> Le 24/02/2013 17:51, Chris Malton a écrit :
>> Hi all,
>>
>> I've been using rrdtool quite successfully for some time.  
>> However..... I've now hit a fairly awkward problem, and can't get my 
>> head around it.
>> https://gist.github.com/cjsoftuk/5024419 is the rrdtool command I'm 
>> running (newlines added by me to make it readable!)
>>
>> The problem is, the "Current" column in the legend is always "0.00 bps".
>> If I switch from the CDEFs to the DEF, it's fine, but then I'm 
>> missing aggregated data.
>>
>> I'm sure I'm being stupid here, however, I can't see where I'm 
>> getting it wrong.
>>
>> Could anyone shed some light on the situation?
>>
>> Thanks in advance.
>>
>> Regards,
>>
>> Chris
>>
>>
>> _______________________________________________
>> rrd-users mailing list
>> rrd-users at lists.oetiker.ch
>> https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20130224/693308e9/attachment-0001.htm 


More information about the rrd-users mailing list