I'm using rrdtool to graph statistics for about 1000 or so interfaces on our network. To get regional information as well as information about our total peering traffic I am generating aggregate graphs from the interface rrd files. However, we ran into a nasty side effect. We keep the rrd files (and the entire polling subsystem etc) in sync with what is actually living on the devices. (i.e. a link is disco'd all the rrd files are removed and the interface is removed from the poller) When this happens we lose all historical data for that link in the aggregate. (The aggregates are being created by a huge CDEF based on the example on the rrdtool web page)
What I'm trying to do is convert over to actually storing the aggregate values in a seperate RRD file. (there will be a process that does all the math before the values get pushed into the file) The first step of this is to convert all the existing data into the new rrdfile. We have 2+ years of data in some cases (converted from mrtg) and I'm running into some problems that I don't know how to approach. The biggest issue is how do I populate the new RRD file? Do I have to look at different time slices for each file to get all of the data? or is there an easier way to do it? i.e. rrdtool fetch AVERAGE -r 300 -s -2y only returns 735 lines. (including the header stuff, just a raw count using wc -l) Changing the -2y to -2d returns 580 rows. Is the answer to run this on all of my RRA boundries stepping forward in time? or is there a better solution. Thanks. - andrew __________________________________________________ Terrorist Attacks on U.S. - How can you help? Donate cash, emergency relief information http://dailynews.yahoo.com/fc/US/Emergency_Information/ -- Unsubscribe mailto:[EMAIL PROTECTED] Help mailto:[EMAIL PROTECTED] Archive http://www.ee.ethz.ch/~slist/rrd-users WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
