[rrd-users] Can i get the <last_ds> value?

[Jayashree]

Hi Alex,

Thanks for your response.
You have described my problem somewhat correctly.

My data availability will be like:-

12:00  counter value = no response
12:05  counter value = 100,300.
12:10 counter value = 10, 600  The difference is 300 bytes in300 seconds = 1
byte per second.
12:15  counter value = 100,900. The difference is 300 bytes in 300 seconds =
1 byte per second.
12:20  counter value = 101,200. The difference is 300 bytes in 300 seconds =
1 byte per second.
12:25 counter value = no response
 12:30 counter value = no response

I am trying to say that,* i will have data running on the interface for a
short duration and the rest of the time there will be no response.*
**
So is setting the heartbeat to a higher value the right way to go about it?

If it helps, I am trying to plot the data on a Data MDT in a multicast
network, where the traffic on the MDT flows only at certain times.
*Option 2 *
**
I have tried the option using 'U'. But here, the problem lies with MRTG
behaviour of "assume last".
It shows me the "currrent in" value at the bottom of the graph  as
 10 bytes @ 12:30  even if there is no traffic on the interface (since the
last valid value was 10bytes @12:20)

To counter that, i used 'unknaszero'. But that also gave me another problem
(sigh!).
With using 'unknaszero' my graph is plotting fine when there is traffic on
the interface, but the legend at the bottom "current in" always is "0
bytes".

So I am still stuck with this problem. :(

Much appreciate any suggestions.


On Thu, Oct 1, 2009 at 3:53 PM, Alex van den Bogaerdt <
alex at vandenbogaerdt.nl> wrote:

>  ----- Original Message -----
> From: "Jayashree" <jayashree.nayak09 at gmail.com>
> To: <rrd-users at lists.oetiker.ch>
> Sent: Thursday, October 01, 2009 12:03 PM
> Subject: [rrd-users] Can i get the <last_ds> value?
>
>
> > Hi,
> >
> > I am interested in getting the counter value stored in <last_ds> in the
> > rrd
> > file. Any rrd function which can help me?
> >
> >
> > The idea is, i dont have response from the router at a certain point and
> I
> > want to log a zero in my graph. So to achieve this, I am trying to get
> the
> > last counter value stored in <last_ds> in my rrd file.  If there is a
> > better
> > way to handle this problem of updating my rrd file with a '0' when i dont
> > have response from the device, please let me know.
>
>
> If you would do that, and if your router would continue working, then at
> the
> next update you will have all of the traffic in the next update interval.
> Is
> that really what you want?
>
> Example:
>
> 12:00  counter value = 100,000
> 12:05  counter value = 100,300. The difference is 300 bytes in 300 seconds
> =
> 1 byte per second.
> 12:10  no response. You update with 100,300. The difference is 0 bytes in
> 300 seconds = 0 bytes per second.
> 12:15  counter value = 100,900. The difference is 600 bytes in 300 seconds
> =
> 2 bytes per second.
> 12:20  counter value = 101,200. The difference is 300 bytes in 300 seconds
> =
> 1 byte per second.
>
> Four intervals of 5 minutes become: 1, 0, 2 and 1 bytes per second.
>
> You can set heartbeat to (some value just over) 600 seconds and do not
> update. Then the example becomes:
>
> 12:00  counter value = 100,000
> 12:05  counter value = 100,300. The difference is 300 bytes in 300 seconds
> =
> 1 byte per second.
> 12:10  no response. You do not update.
> 12:15  counter value = 100,900. The difference is 600 bytes in 600 seconds
> =
> 1 byte per second.
> 12:20  counter value = 101,200. The difference is 300 bytes in 300 seconds
> =
> 1 byte per second.
>
> Four intervals of 5 minutes become: 1, 1, 1 and 1 bytes per second.
>
>
> You can also signal that you don't know what happened by inserting an "U"
> in
> the data:
>
> 12:00  counter value = 100,000
> 12:05  counter value = 100,300. The difference is 300 bytes in 300 seconds
> =
> 1 byte per second.
> 12:10  no response. You update with "U". The difference is unknown so the
> rate becomes unknown.
> 12:15  counter value = 100,900. The difference is unknown so the rate
> becomes unknown.
> 12:20  counter value = 101,200. The difference is 300 bytes in 300 seconds
> =
> 1 byte per second.
>
> Four intervals of 5 minutes become: 1, unknown, unknown, 1 bytes per second
>
>
> If the unavailability of your router lasts longer than the heartbeat
> interval time, then above is what automatically would happen, even if you
> do
> update with a known counter value.
>
>
> HTH
> Alex
>
> _______________________________________________
> 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/20091005/7e795c2d/attachment.htm