[rrd-users] implicit generation of HW RRAs differ from the description

[Norbert Preining]

Dear all,

I was checking how to tweak the HW prediction and started to switch from
implicit generation of the various RRA to explicit.

It turned out that what the rrdtool create does, and what is described
in the manual page (and also is close to my intuition), differ
considerably.

Please consider the following test script:

alpha=0.4
beta=0.001
start=$(date --date="2016-09-01 00:00" +%s)
commonargs="--start $start --step 5s DS:someval:GAUGE:20m:U:U RRA:AVERAGE:0.5:5s:3d"
rm auto.rrd manual.rrd
rrdtool create auto.rrd	$commonargs	\
	RRA:HWPREDICT:1w:$alpha:$beta:1d

rrdtool create manual.rrd $commonargs	\
	RRA:HWPREDICT:1w:$alpha:$beta:1d:3 \
        RRA:SEASONAL:1:$alpha:2 \
        RRA:DEVSEASONAL:1:$alpha:2 \
        RRA:DEVPREDICT:1w:4 \
        RRA:FAILURES:1:7:9:4

rrdinfo auto.rrd > auto.info
rrdinfo manual.rrd > manual.info

diff -u auto.info manual.info

exit 0

The above two invocations generate the same content checked via rrdtool dump and diff
(only differences are in cur_row values).

Now this seems to be in contrast to the description in the man page of rrdcreate:
       The SEASONAL and DEVSEASONAL RRAs store the seasonal coefficients for the Holt-Winters
       forecasting algorithm and the seasonal deviations, respectively.  There is one entry per
       observation time point in the seasonal cycle. For example, if primary data points are
       generated every five minutes and the seasonal cycle is 1 day, both SEASONAL and
       DEVSEASONAL will have 288 rows.
But indeed, the dump of the rrdtool auto creation does set the seasonal and devseasonal to 1 row only!

Furthermore, concerning the RRA:FAILURES:
       argument. If the FAILURES RRA is implicitly created, rows will be set to the seasonal
       period argument of the HWPREDICT RRA. Of course, the RRDtool resize command is available
       if these defaults are not sufficient and the creator wishes to avoid explicit creations
       of the other specialized function RRAs.
Furthermore, also the failures only contain 1 row, which is incorrect.


So if I want to create something that conforms with the description in the man page,
I need to use
rrdtool create manual.rrd $commonargs	\
	RRA:HWPREDICT:1w:$alpha:$beta:1d:3 \
        RRA:SEASONAL:1d:$alpha:2 \
        RRA:DEVSEASONAL:1d:$alpha:2 \
        RRA:DEVPREDICT:1w:4 \
        RRA:FAILURES:1d:7:9:4


Now the question is, which data does the internal rrd HW computation use? 
Is there a misunderstanding from my side, or is something missing in my view?


Thanks a lot

Norbert

--
PREINING Norbert                               http://www.preining.info
Accelia Inc.     +    JAIST     +    TeX Live     +    Debian Developer
GPG: 0x860CDC13   fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13