Cheshire Cat Computing

The first [_] sets up the default for Target[]s without an explicit options string, so things get the 'active'

The second overrides this for the to-sw01_0 target, so this taregt doesnt get 'active'

The userdefined graph then inherits from the to-sw01_0 target and doesnt get 'active'.

# routers.cgi*Options[to-sw01_01IE]: none
# routers.cgi*Options[to-sw01_01I_IE]: none

these two would explicitly set the options for the specified userdefined graphs, else they'll inherit the options from their respective primary targets, which (unless explicitly set as above) will inherit from the [_] target.

Using the routers.cgi*Options[_] twice in a file will probably only result in the last one taking effect. If you set the default twice to different values its undefined which will take effect, really. routers2 applies defaults after the file has been completely read, although I think that now MRTG applices them at some undetermined step during the processing.

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

Hello Steve , Then it looks like I'll have to scan thru all the interfaces OR not see the statistics , until some show up that is .
To be honest I'd really like to see all the stats than wait for an error to appear .
But I have users that get confused by empty graphs . Not a disparagement of them , just a fact .
So I'll probably go with the blanket 'routers.cgi*Options[_]:active' just to keep the confusion down .

But it sure would be nice to be able to use a regexp in that field some how .
Twyl , JimL

I can see your point. The trouble is, it would be extremely difficult to do with the existing code structure, not to mention that it would diverge even further from the MRTG .cfg file standards (we already diverge by allowing multiple routers.cgi*Graph, *Summary, *Link and *Extension directives). I think that it will have to wait until a later version, sorry...

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

Hello Steve & All , Along the same lines .
is there a way to have the ...
In Errors: Max -0.00 bps, Avg -23.20 bps, Last -0.00 bps
Out Errors: Max -0.00 bps, Avg -0.00 bps, Last -0.00 bps

have a differant color ?
and Maybe In both the text based & Graphic ?
Tia , babydr(JimL)

This you can do. Either put a list of colour codes on the first routers.cgi*Graph[] definition, or else use routers.cgi*Colours[userdefinedgraphname]: followed by a list of colour codes. The colour codes are used in rotation (IE, if you dont give enough, it restarts from the beginning of the list again). With the routers.cgi* prefix it can take more than 4 options (the standard MRTG directive can take only 4 colours).

The colour codes must be 6 hex digits with a leading '#'. Eg, the default for userdefined graphs in colour mode is:

routers.cgi*Colours[userdefinedgraphname]: #00ff00 #0000ff #ff0000 #00cccc #cccc00 #cc00cc #8800ff #88ff00 #ff8800 #0088ff #ff0088 #00ff88

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

Hello Steve , Using the same example config from above . If I do ...
...
# Colours[to-rtr01_1]: RED1#FF0915,RED1#FF0915,RED2#CC000B,RED2#CC000B < Not wanted here
...
Colours[to-rtr01_1IE]: RED1#FF0915,RED1#FF0915,RED2#CC000B,RED2#CC000B
... and/or
Colours[to-rtr01_1I_IE]: RED1#FF0915,RED1#FF0915,RED2#CC000B,RED2#CC000B
...

The text output in summary mode the above had no effect on . In the graph itself there was no change either , I'll try uploading the graphic to here , But last time I tried I was unable to check that it was succussful or not ie: it never showed anywhere . Well it made it , yay !
You will note in the graph that the IIE/sec & the OIE/sec show default colours , not the ones I Hopefully selected .
Tia , babydr(JimL)

I notice you're using HRule for the grey lines, remember that you can also have
routers.cgi*HRule[graphname]: 64000 "Maximum bandwidth"
if you want the legend to have a label. The HRule lines currently are all in (automatic) shades of grey, though, as you can't give a colour code to the HRule directive - this is something I should really add in the next version.

The Colours[] directive does not affect the text colour on the page, only in the graph image. Sorry if I misunderstood you previously. The summary/compact pages will always use blue and green.

In the graph image, you are still getting the default colour set, which means the colours are not getting loaded. Make sure you have the routers.cgi*Colours[] line at the end (not before the first routers.cgi*Graph[] definition, otherwise the userdefined graph will not yet be defined for the option to take effect) and that the graphname is correct.

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

routers.cgi*Graph[to-rtr01_1]: to-rtr01_1I-IE
routers.cgi*Graph[to-rtr01_1IE]: to-rtr01_1I-IE
routers.cgi*InMenu[to-rtr01_1I-IE]: Yes
routers.cgi*InSummary[to-rtr01_1I-IE]: Yes
routers.cgi*Desc[to-rtr01_1I-IE]: 01 1 To: MO network , Serial1/0 Traffic & Error Rates IN/OUT
routers.cgi*HRule[to-rtr01_1I-IE]: 64000 ""
routers.cgi*HRule[to-rtr01_1I-IE]: 256000 ""
routers.cgi*Colours[to-rtr01_1I_IE]: RED1#FF0915,RED1#FF0915,RED2#CC000B,RED2#CC000B

Here is a whole Userdefined section .
Please show me where I am going wrong here .
Fyi , What I am trying to do here is make the Error Graph & statistics very noticable . Maybe with that said another idea might crop up ? failing that is there a place I could possibly hack in an optional 'text' colouration ? Hints ?
Tia , babydr(JimL)

[quote]Post subject:
PostPosted: Sun 08 Aug, 2004 11:58 am

Joined: Mon 28 Jul, 2003 1:42 pm
Posts: 2392
Location: Auckland, New Zealand
This is not actually a feature of rrdtool as such, what the author has done is to change the background colour according to a second value. This is done by createing several pseudo-values, in the same way as I make the 'working hours' background go yellow or the 'unknown' go grey.

However this is intriguing -- maybe I could make the background go red if the value or either in or out was outside the defined thresholds, for example? Or even make the in/out line/block change between three (or more) different colours depending on a userdefined range, as you suggest -- although this would slow graph generation considerably.

I also have plans for a 'mirror' graph style, where in is shown above the axis, and out below the axis (both as an area). This would allow the combined incoming and outgoing to be shown stacked together on the same graph. This is being looked at for v2.15/2.16.[/quote]

Wow now this would be Great for making the Error Graphs be quite Obvious there's a difficulty . Did this ever get into production ? If so please how to do .
Tia , babydr(JimL)

The mirror graph mode is available. I've never yet set up the one with the background going red if the values are outside of the thresholds, or outside of the Holt-Winters predicted range (now *that* would be cool, although complex). Multiple ranges to change colours would probably be too complex for now.

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