I've come to rrd via ntop (http://www.ntop.org/) which I'm looking at in the context of sFlow (http://www.sflow.org/) and receiving sFlow counter samples as often as once per second per interface. RRD and its use by ntop seems rather ideal in that it has the concept of keeping so many samples at fine resolution, so many at a coarser and so on. I've (or at least I think I have) cranked ntop's storage of values into RRD down to once per second, but the appearance of messages such as:
> Mon Apr 18 15:06:07 2011 **WARNING** RRD: > rrd_update(/usr/local/var/ntop/rrd/interfaces/sFlow-device.2/sFlow/48/ifOutMulticastPkts.rrd) > error: > /usr/local/var/ntop/rrd/interfaces/sFlow-device.2/sFlow/48/ifOutMulticastPkts.rrd: > illegal attempt to update using time 1303164367 when last update time > is 1303164367 (minimum one second step) (Which may have appeared even before the rate was cranked down to every second, I must confess I'm not quite sure one way or the other) has me concerned about operations at the limit - in other words, ostensibly I have values going into the rrd once per second, but suppose that flutters back and forth across the second boundary? Is it likely that is what happened above? That the update happened to hit just before the second rolled-over? In this instance I happen to be running 4.0.3 of ntop compiled from source, and: raj@tardy:~$ dpkg -l | grep rrd ii librrd-dev 1.4.3-1ubuntu1 time-series data storage and display system (development) ii librrd4 1.4.3-1ubuntu1 time-series data storage and display system (runtime library) ii rrdtool 1.4.3-1ubuntu1 time-series data storage and display system (programs) for the RRD bits. thanks, rick jones _______________________________________________ rrd-users mailing list rrd-users@lists.oetiker.ch https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
