> From: Bert Driehuis
> I've had an unrelated cause for spikes that render the charts of the
> history of my Squid server useless. RRDtool (at least in the way it is
> called by Cricket, but I believe this to be generally true) cannot cope
> with a restart of an SNMP agent. COUNTER objects wind up with huge
> values in this scenario:
>
> Time Value PDP Value
> t 200 0
> t+1 250 50
> t+2 300 50
> [ agent restarts, counter goes to zero ]
> t+3 5 4.2e9
I just dealt with this kinda problem yesterday: the cure for
me/nrg is to set the min and max values for each DS... and I
happen to have handy examples laying around...
Try
rrdtool create -b 947517820 test.rrd \
DS:queries:COUNTER:600:U:U \
RRA:AVERAGE:0.5:1:10
vs
rrdtool create -b 947517820 test.rrd \
DS:queries:COUNTER:600:0:1000 \
RRA:AVERAGE:0.5:1:10
with the data
rrdtool update test.rrd -t queries 947517827:26802974
rrdtool update test.rrd -t queries 947518294:26820088
rrdtool update test.rrd -t queries 947518614:451
rrdtool update test.rrd -t queries 947518867:6970
rrdtool update test.rrd -t queries 947519166:6970
The later doesn't spike because its sets the DS max value.
But it assumes a max rate of 1000 queries/sec.
If we assume this data is traffic at OC3 rates and set the the
max to 19375000 (1.9e+07 aka 155 mbits/sec) I still get a spike
in the 1+e06 range. Which makes sense, I think, because 1+e06
is less than 1.9e+07.
But is there a graceful way to handle the example above with
19375000 as the max??
later
steve
- - -
systems guy
wiscnet.net
--
Unsubscribe mailto:[EMAIL PROTECTED]
Help mailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-developers
