Cheshire Cat Computing

Hello all ,
I am using Factor[name]: -1 , Well actually alot larger than that but even -1 gives my
Maximums as '0.00' .
Am I going something wrong here , Or might there be a work arround ?
Tia , JimL

I'll have to look into this - I've never really tested with negative factors, so there might be some things in the code that assume >0. Also, by using a -ve factor, you've actually turned it into a minimum, so that might affect some things in the RRD library, too.

Can you post (or email) a snip of the .cfg file, and a copy of the generated graph image? I cant promise to be able to work on it immediately, but this will definitely help.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hello Steve , It's in the mail . Subject is the same as here .
hth , & tia , JimL

Hello Steve , Sent a reply via email . Included below in case of failed comm.s .

On Tue, 8 Apr 2008, Steve Shipway wrote:
> Thanks for this - I can see what's happening.
>
> Basically, the RRD stores the max value of the metric. However, when
> the Factor is negative, it means that what we actually want is the max
> value of the *absolute* value (or the minimum value of the factored
> value). This confuses the system because routers2 always asks for a
> maxima, but in this case we need a minima. However, if I get routers2
> to ask for a minima when it sees a negative factor, then this confuses
> RRD which realises it doesn't have any minima RRAs available. It may be
> better under later versions of RRDTool, though.
Ok I see the problem . Not sure of the later releases of rrd . Ours is
# rrdtool --version
RRDtool 1.2.26 Copyright 1997-2007 by Tobias Oetiker <tobi@oetiker.ch>
Compiled Feb 22 2008 10:01:43

> Would you be willing to run some tests for me? If I can send you a
> v2.18 beta with some code to try and sort out max/min, maybe you could
> try it out and let me know if it acts sane? Otherwise I'll need to wait
> until I get time to work on this, and things are very busy around here
> at the moment.
No problem , send away , I can sure try here .
Tnx , JimL

Maybe already mentioned before .

the next cavaet is that the ...
MaxBytes[mo-rtr01_1]: 8000
AbsMax[mo-rtr01_1]: 32000

Does not plot their lines . and/or the 'Custom Graph' doesn't allow them being plotted ?

A Question tho is there a way to have a line at a particular + or - value to be plotted as long as teh graphed data goes that far ?

Fyi , see "Not all data being displayed . Config inside ." in this forum .

Tia , JimL

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hello Steve , While the work around gives me a line to check against (Thank You) it isn't the same as the real thing (tm;-) .
And the discussion above leads me to beleive that there is/was a misconception of the area the 'MaxBytes' & 'AbsMax' entries appear .

... # In the true Positive only BW section .
MaxBytes[mo-rtr01_1]: 8000
AbsMax[mo-rtr01_1]: 32000
... # In the error rate ie: "Factor[to-rtr01_1IE]: -1000000" section .
MaxBytes[to-rtr01_1IE]: 1
... # The user define area .
No MaxBytes or AbsMax in the "routers.cgi.*[to-rtr01_1I-IE]:" area .

With them arranged as above in the test sample I sent to
"Post subject: Not all data being displayed . Config inside ."
on "Posted: Thu 10 Apr, 2008 3:35 pm AKT"
One would hope that the section that is '+' would at least plot those lines .
The other two areas I (at present) am not so worried about , Tho others might .
Tia , JimL

Ah - MaxBytes and AbsMax are not inherited by a userdefined graph, because each sub-Target may have a different value. In this case, it would need to plot one Max line for each target (since otherwise there would be no way to know which was which). If course, if all MaxBytes values were the same, then it would work, but that would need to be specifically checked for.

I can't remember if setting routers.cgi*MaxBytes[userdefinedgraphname]: will give you a Max line specific to the userdefined graph, but I suspect not. This would probably be the 'correct' way to do it, though.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

You are correct that the 'routers.cgi*MaxBytes[userdefinedgraphname]:' and the 'routers.cgi*AbsMax[userdefinedgraphname]:' do not plot their respective lines on a user defined graph , at least not in the graph setup I have .
thank you for the idea .

I've just checked the code.

While routers.cgi*AbsMax and MaxBytes can be defined for userdefined graphs, the userdefined graph code does not actually ever display Max or AbsMax lines. I think the logic behind this was that it doesnt make sense to have this for combined graphs usually, and if you want to have a Max line, then you should use routers.cgi*HRule[] to add one.

It may be worth making routers.cgi*AbsMax/MaxBytes get supported for userdefined graphs in the future, but I dont think they should ever be inherited from the primary target.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11