[rrd-users] Wrapping not working correctly

[Lordy]

Hello,

I just started using rrdtool actively for graphing my network usage a few 
days ago.

So far all has been running pretty smooth _except_ for the counter wraps. 
Here is
a short description of what I do:

I have a perl script that collects counter values from /proc/net/dev (Linux 
2.4.20)
every 60 seconds. Those counters are then put into an rrdfile that I 
created with this command:

rrdtool create eth0.rrd DS:input:COUNTER:3000:U:U DS:output:COUNTER:3000:U:U
RRA:AVERAGE:0.5:1:3000 RRA:AVERAGE:0.5:30:3500 RRA:AVERAGE:0.5:120:4000
RRA:AVERAGE:0.5:1440:4000 RRA:MAX:0.5:1:3000 RRA:MAX:0.5:30:3500
RRA:MAX:0.5:120:4000 RRA:MAX:0.5:1440:4000

This is a log of the counter wrap and the according values from the rrdfile:

TIMESTAMP : INPUT : OUTPUT (log)
1052944620 : 26649774336 : 34299525336
1052944680 : 26653861104 : 83380896
1052944740 : 26657344136 : 214937496
[...]
1052931120 : 25223793736 : 34297051328
1052931181 : 25229329384 : 150880920
1052931240 : 25234327032 : 343693456

TIMESTAMP : INPUT(AVERAGE) : OUTPUT(AVERAGE) (rrd)
1052944500: 3.5023817268e+04 1.1066858514e+06
1052944800: 6.5019546667e+04 6.1489146800e+16
1052945100: 9.9007487213e+04 3.0024917578e+06
[...]
1052931000: 9.3047440000e+04 3.2095458133e+06
1052931300: 8.1707626667e+04 6.1489146801e+16
1052931600: 7.2381360000e+04 2.9875407733e+06

Is this a problem with /proc/net/dev, rrdtool or just sh!t in -> sh!t out ?

I'm really curious to hear any comments on this as rrdtool is supposed to 
"handle wraps
gracefully" according to Tobi's earlier postings on this list.

Best regards,
Lordy 

--
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
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi