[rrd-users] Re: What am I doing wrong?
Serge Maandag
serge.maandag at staff.zeelandnet.nl
Wed Oct 15 15:46:15 MEST 2003
Changing the xff will not help you here.
I believe it works like this:
Say you update 5 times within the step size time.
The updates have the values (1, nan, 3, nan, nan).
An RRA MAX:0.9999999999:1:9999 would get the value 3.
An RRA MAX:0.5:1:9999 would get the value nan, because more than half of
the values are nan.
The second RRA (which consolidates more than 1 pdp) considers it's xff
when values from the first RRA are consolidated into it.
In your case your updates have the values (nan, nan, nan, nan, nan)
because everytime you update, you are well beyond the heartbeat of the
DS.
In other words: Fiddling with the xff won't help you because damage has
already been done.
Choose a step that is close to the average time your updates and choose
a heartbeat that is beyond the biggest interval between updates.
Serge.
-----Original Message-----
From: Dean Takemori [mailto:dtakemori at home.comtelhi.com]
Sent: Tuesday, October 14, 2003 11:28 PM
To: Serge Maandag; rrd-users at list.ee.ethz.ch
Subject: Re: [rrd-users] What am I doing wrong?
Thank you for the prompt reply.
Unfortunately, I must STILL be misunderstanding something.
My reading of the documentation suggests that the following data
1066093663:0:1:0:0
1066096371:0:1:0:0
1066097939:0:1:0:0
1066155010:0:1:0:0
1066155844:0:1:0:0
1066160322:0:1:0:0
with "RRA:MAX:0.9999999999:1:9999
1066089300: nan nan nan nan
1066089590: nan nan nan nan
1066089880: nan nan nan nan
[snip all nan's across the board]
1066164990: nan nan nan nan
1066165280: nan nan nan nan
1066165570: nan nan nan nan
Should have some non-nan entries, since I am basically asking to return
the MAX value in a "step" if any single value is defined.
--
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