[rrd-users] Graph Never goes over 100M
Big Wave Dave
bigwavedave at gmail.com
Wed Oct 1 19:25:43 CEST 2008
On Wed, Oct 1, 2008 at 10:09 AM, Chris Fogel <chris.fogel at intermetro.net> wrote:
>
> I have read the lists and googled, but cannot figure out why rrdtool
> graphs all values over 100M incorrectly.I have checked my .rrd files and
> the values in the db are the correct values.
>
> In this example you will see at 8:30am, the traffic goes to 145M, but is
> graphing at 45M instead, and does this for all values over 100M, then at
> ~14:00, the value drops back down under 100M and it graphs normally again.
>
> Below is the code used to generate it...
>
> Thanks,
> Chris
>
> #!/bin/bash
> rrdtool graph /var/www/localhost/htdocs/development/pngs/day/$1.png \
> --rigid \
> --upper-limit 200000000 \
> --title "Port Bandwidth" \
> --interlace --imgformat PNG \
> --width=700 --height=150 \
> --start "now -16h" \
> --watermark "`date`" \
> -v Bandwidth \
> DEF:a=/var/www/localhost/htdocs/development/rrds/$2/$1.rrd:ds0:AVERAGE \
> DEF:b=/var/www/localhost/htdocs/development/rrds/$2/$1.rrd:ds1:AVERAGE \
> CDEF:alpha=8,a,* \
> CDEF:beta=8,b,* \
> CDEF:zeta=a,8,*,b,8,*,GT,a,8,*,b,8,*,IF,PREV,GT,a,8,*,b,8,*,GT,a,8,*,b,8,*,IF,PREV,IF
> \
> COMMENT:"\t avg. max last \n" \
> AREA:zeta#aaaa55:Tot \
> GPRINT:zeta:AVERAGE:" %4.2lf %s" \
> GPRINT:zeta:MAX:"%4.2lf %s" \
> GPRINT:zeta:LAST:"%4.2lf %s\n" \
> LINE1:beta#ff0000:Out \
> GPRINT:beta:AVERAGE:" %4.2lf %s" \
> GPRINT:beta:MAX:"%4.2lf %s" \
> GPRINT:beta:LAST:"%4.2lf %s\n" \
> LINE1:alpha#0000ff:In. \
> GPRINT:alpha:AVERAGE:" %4.2lf %s" \
> GPRINT:alpha:MAX:"%4.2lf %s" \
> GPRINT:alpha:LAST:"%4.2lf %s\n"
>
> --
> Chris Fogel, InterMetro Communications, Inc.
> 805.433.8000/o, 805.433.0049/d, 805.582.1006/f
> chris.fogel at intermetro.net, http://www.InterMetro.Net
>
> _______________________________________________
> rrd-users mailing list
> rrd-users at lists.oetiker.ch
> https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
>
>
I have seen this when graphing a 32-bit SNMP value that "wraps" when
it exceeds a 32-bit size. I ended up having to change the SNMP
queries to use a 64-bit counter.
Dave
More information about the rrd-users
mailing list