[rrd-users] Re: Cisco NAM + Perl + SNMP + RMON2 + RRDtool
serge.maandag at staff.zeelandnet.nl
Wed Dec 8 23:11:07 MET 2004
> I was attempting to store all 8
> values in one RRD but am wondering if instead I'll need to
> use multiple ones.
If you stick multiple DS's in a RRD database, you'll have
To update them all at once. You cannot update half and update
The rest in another go.
Also, adding or extracting DS's from a RRD database is a messy
Job that you'd rather avoid.
If neither of these 2 problems will affect your case, then you're
Safe to use one single database. The benefits of using one database
Are uhm.. Less files on your disk, so less slack. You need less
Filehandles and perhaps it's a little faster.
> like an MRTG/14all style page that shows the last 24 hours,
> as well as a
> weekly, monthly, & yearly average. I created the RRD as follows:
> /usr/bin/rrdtool create /tmp/dfw-core1-nam-packets.rrd -s 60
> DS:SMTP:GAUGE:60:0:U DS:POP3:GAUGE:60:0:U DS:HTTP:GAUGE:60:0:U
> DS:LEFTOVER:GAUGE:60:0:U DS:SMTP-OVERFLOW:GAUGE:60:0:U
> DS:POP3-OVERFLOW:GAUGE:60:0:U DS:HTTP-OVERFLOW:GAUGE:60:0:U
> DS:LEFTOVER-OVERFLOW:GAUGE:60:0:U RRA:LAST:0.5:1:4000
Your step is 60 seconds and your heartbeat is too. Usually
You'll want to set the heartbeat a bit larger than the step size,
That is two times the same resolution, but with different sizes.
The first one is redundant.
> RRA:LAST:0.5:1440:800 RRA:MAX:0.5:1:4000
> RRA:MAX:0.5:30:800 RRA:MAX:0.5:120:800 RRA:MAX:0.5:1440:800
Same here for the MAX function
> 1) Why is the data stored one polling cycle behind? Is that because it
> gets stored in the last PDP field and tagged on in the next cycle?
Rrdtool internally works with rates. If you do an update it looks if
You have past the row time for every RRA. If so, it calculates
Averages, maxes and so on and pushes them on top of the RRAs.
That's why you are always one row time behind. In your case that's
60 seconds for the highest res RRAs
> 2) Why are the values not matching in all cases?
I just realized that the LAST function isn't all too clear to me.
It doesn't seem to be explained on the rrdtool site, but I may have
> 3) Should I actually be storing the values in one RRD that I use to
> calculate the deltas and then storing those in another RRD
> that I graph,
> instead of trying to cram it all into one RRD?
Uhm, already answered that one :)
Op de inhoud van dit e-mailbericht en de daaraan gehechte bijlagen is de inhoud van de volgende disclaimer van toepassing: http://www.zeelandnet.nl/disclaimer.php
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-users-request at list.ee.ethz.ch?subject=help
More information about the rrd-users