[rrd-users] Re: normalisation
thomas at braindumped.com
Thu Apr 27 13:11:38 MEST 2006
in my last mail i only replyed to you. not to the list. sorry.
but i still got a question.
Am 26.4.2006 schrieb "Alex van den Bogaerdt" <alex at ergens.op.het.net>:
>On Wed, Apr 26, 2006 at 04:57:47PM +0200, thomas peter wrote:
>> and start updating at 1146061700 and then update at 1146062000 and at
>> 1146062300 and so on, i can dump the exact values i provided to the
>1146061700 = 3820205 * 300 + 200
>1146062000 = 3820206 * 300 + 200
>and so on.
ok, bad example. rrdtool makes shure that the slots are multiples of 300
i 'grep'-ed for 'last_update' in the rrdinfo output to determine the
value i base my calculations on.
but when i once introduced an offset in updateing, i won't get rid of
it, until i change the plan to gather the correct slice-time.
if i knew the --start parameter that was provided during rrdcreate it
would be easy. but how do i get the value out of an existing database?
really grep for an occurence of <row>, get the unixtimestamp out of that
line and calculate the last valid slot from now? (or better the next
valid slot and sleep until that time has come [and than gather fresh
data to fill in])
rrdfirst accepts an parameter named rraindex.
do i have to look in a rrddump to gather a valid rraindex or is there a
that rraindex would serve my calculation perfectly well, wouldn't it?
>You need to have updatetime = n * 300
yes. the last 20 lines of this mail i tried to question how to determine
can you give me a hint, if i don't have the information of rrdcreate?
thanx in advance.
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-users-request at list.ee.ethz.ch?subject=help
More information about the rrd-users