[rrd-users] Re: RRds::graph

Marco Flohrer marco.flohrer at informatik.tu-chemnitz.de
Tue Jul 3 00:43:00 MEST 2001

On Mon, 2 Jul 2001, Alex van den Bogaerdt wrote:

> Marco Flohrer wrote:
> > If i define minimal and maximal value for an data source,
> > *UNKNOW* will be written to the database.
> > But also time of last_update is changed to actual time.
> Yes, an update is an update, even if you are updating with unknown.
> > I want that in such a case time of last_update is _not_ changed and
> > ds[In].last_ds is also untouched.
> > My RRD's have only one data source.
> > Is this in any form possible or have i to call 'rrdtool update' only if
> > minimal and maximal value are OK and test this by myself?

This is my current solution.

> That would probably be best.  However, maybe there's another solution
> if you let us know your problem.

Our switches (3com 3300, rmon hostTable-oid's) are slow in actualizing
some conters. Depending in current traffic, time between last and next
actualization can be under 1 minute or greater 7 minutes. For good
graphing results more then 5 minutes base interval (--step) is no option.
So base intervall is 5 minutes, i get actual counters every 1 minute from
all switches. The mentioned counters are most times not different from
it's predecessors, so not written to the RRD.

This is not perfect: when i get data via SNMP i do not know _when_ was
it's last actualization, so graph looks a bit like a wave...
I could live with only the "broken" data sources to store them in another
way: 10 minutes base intervall or more than 1 primary data point for one
data point in first RRA. But in near future it seems that we have also
for this data sources precise counters, so "normal" base intervalls
etc. are possible again. Is there any way to convert an exisiting RRD to
a new one, for example

 --steps 300 RRA:AVERAGE:0.5:2:rows RRA:AVERAGE:0.5:6:rows

 --steps 300 RRA:AVERAGE:0.5:1:rows RRA:AVERAGE:0.5:6:rows


Data loss in first RRA ist no problem, looks only a few days a bit

mailto: marco.flohrer at informatik.tu-chemnitz.de
talk:   mafl at diamond.csn.tu-chemnitz.de

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

More information about the rrd-users mailing list