[rrd-developers] Re: Data Loss on non-contiguous updates

Lee Thompson bm55b at yahoo.com
Mon Aug 15 19:35:56 MEST 2005


>--- Tobias Oetiker <oetiker at ee.ethz.ch> wrote:
>
> Lee,
> 
> you are no particularly fond of the
> minimal-required-heartbeat
> option I gather ?

I didn't like it or dislike it, just didn't understand
it.  Thanks for your expanation.  I would think its
not an option but a requirement as I can't figure out
how to set mrhb to 1.  In the perl program I sent, I
had 

  RRDs::create ($rrd, "--start",$start-1,
"--step",300,
              "DS:a:GAUGE:300:U:U",
              "RRA:AVERAGE:0.99999:1:300",
              "RRA:MIN:0.5:12:300",
              "RRA:MAX:0.5:12:300",
  );

So with a heartbeat of 1 sample (300), I was hoping
RRD would record the first sample.  Changing the
heartbeat to 0 or 1 didn't help.  Neither did setting
the RRA:AVERAGE to RRA:LAST.

To preserve the first sample, is writting udefined
(NaN) to the prior sample the proper way of dealing
with this issue given RRD's design or is there a way
to change the RRDs::create(); to collect the first
sample?



While I can understand Alex's statement that this
belongs on the user list, I hope you can appreciate
that the docs only talk about the impertive for
contiguous sampling and does not define the edge case
for skipped records and re-basing the start as Tobi
described.  This left me trolling around the source
code looking for answers as I thought I was looking at
a bug, not a design decision.  From my point of view,
all I knew was I was missing data - a very bad thing. 
Some googling on rrd email lists didn't get me a hit.

A good ehancement would be an extension to the
rrd-beginners docs for programatic examples on how to
program the restart of collector deamons with an eye
towards preserving samples.  

Thanks for input,
Lee Thompson




		
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs 
 

--
Unsubscribe mailto:rrd-developers-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-developers-request at list.ee.ethz.ch?subject=help
Archive     http://lists.ee.ethz.ch/rrd-developers
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi



More information about the rrd-developers mailing list