[drraw-users] Unable to graph with VDEF operators LSLSLOPE, LSLINT, and LSLCORREL

[Nokes, Sam X]

Appreciate the assist David. Unfortunately I received the familiar "rpn
expressions without DEF or CDEF variables are not supported" error
 
Here is the rrd invocation generated by drraw for review:
rrdtool graph - \
 --start=end-86400 \
 --title=CPU Projection \
 --vertical-label=Utilization \
 --imgformat=PNG \
 --width=600 \
 --base=1000 \
 --height=200 \
 --interlaced \
 
DEF:a=/data/nmon2web/html/020256D3D/Shared/c0prd01/rrddata//cpu_all.rrd:
UserPct:AVERAGE \
 
DEF:b=/data/nmon2web/html/020256D3D/Shared/c0prd01/rrddata//cpu_all.rrd:
SysPct:AVERAGE \
 
DEF:c=/data/nmon2web/html/020256D3D/Shared/c0prd01/rrddata//cpu_all.rrd:
WaitPct:AVERAGE \
 CDEF:G=a,b,c,+,+ \
 VDEF:D=G,LSLSLOPE \
 VDEF:H=G,LSLINT \
 VDEF:F=G,LSLCORREL \
 CDEF:E=D,COUNT,*,H,+ \
 LINE1:G#FFA500: Usr+Sys+Wait \
 LINE1:E#FF0000: Projection

________________________________

From: drraw-users-bounces at taranis.org
[mailto:drraw-users-bounces at taranis.org] On Behalf Of David Nolan
Sent: Friday, July 20, 2007 2:35 PM
To: drraw-users at taranis.org
Subject: Re: [drraw-users] Unable to graph with VDEF operators
LSLSLOPE,LSLINT, and LSLCORREL




On 7/19/07, Nokes, Sam X <nokess at upmc.edu> wrote 

	Realizing v2.2a3 is beta, and noting a comment by Alex: "In
RRDtool 1.2 I made a start with VDEF processing. VDEFs aren't arrays
like CDEFs are, they are single values. The work is far from finished,
however certain functions are already usable."(
http://www.vandenbogaerdt.nl/rrdtool/total.php  ), the questions follow:
	(1) Are the VDEF operators LSLSLOPE, LSLINT, and LSLCORREL
usable in drraw v2.2a3?
	(2) Should the vnames be placed into a linear equation:
y=LSLSLOPE*LSLCORREL+LSLINT ?
	   We tried to add a CDEF definition substituting VDEF vnames in
a vpn format (D,F,*,H,+) and 
	   received an error from drraw stating "rpn expressions without
DEF or CDEF variables are not supported"
	(3) Has anyone generated a least squares projection with rrdtool
similar to the following? If so what is the rrd invocation?


Sam, 


I've never used these functions in rrdtool before, but my reading of the
documentation indicates that you are close with statement 2 above.  
I believe you need to define the VDEF variables as you're doing, and
then define a CDEF using those variables and the automatic
COUNT variable, i.e. CDEF:trend=D,COUNT,*,H,+  (using the variables you
defined above), and then graph the CDEF. 

-David




-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://web.taranis.org/pipermail/drraw-users/attachments/20070720/2c171d88/attachment.html