[rrd-users] Re: Help with spikes

[Jon Rust]

At 11:49 PM +0100 3/8/00, Tobias Oetiker wrote:
>
>in that case only Alexs idea with the missing MAX setting is an
>option ... otherwhise we got another bug at our hands ... if so I
>would apreciate if you could provide a script to reproduce this
>behaviour ...
>
>cheers
>tobi

Yes, I'll check the max settings (see below). As far as reproducing 
it, I can usually get it to go by monitoring a port that gets 
shutdown for 1 or 2 pollings. Ports being polled are Cisco 7206 
serial ports and sub-ints, and Cisco 2924 10/100's.

I never had this problem with plain ol' mrtg. Let's check the limit values...

In fact, there aren't any limits per Alex van den Bogaerdt's assersion:

    mail:upstreams{31} $ rrdtool tune totalup.rrd
    DS[ds0] typ: COUNTER    hbt: 86400      min: NaN        max: NaN
    DS[ds1] typ: COUNTER    hbt: 86400      min: NaN        max: NaN

I just created a new entry in my mrtg.cfg file like so:

    Target[test]: -12:public at c2924
    MaxBytes[test]: 1250000
    Title[test]: test rrd
    PageTop[test]: test rrd
    Directory[test]: colos

Let it run, then checked the rrd file:

    mail:colos{37} $ rrdtool tune test.rrd
    DS[ds0] typ: COUNTER    hbt: 600        min: NaN        max: NaN
    DS[ds1] typ: COUNTER    hbt: 600        min: NaN        max: NaN

So is setting a max value something I should do manually on every 
mrtg+rrd created file? And I can safely add the max values at any 
time?

jon

--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users