Stefan M. Brandl wrote: > > Timezone ? >> > >Seems not to be the problem. >If I set the start-date 3600 seconds higher or 3600 seconds lower, >the result is always the same. Startdate is always set to 1267056000 (look >in the resulting rrd-dump). > >> Your email says +2 (European summer time ?), so midnight by your >> clock is 22:00 UTC. RRCTool only uses UTC. >> >> In any case, the start time in an RRD file is largely irrelevant. > >Why? >I want a 1 week interval with data inserted every friday.
I'd missed the step size of 604800. Because that's the way it works<period>. Every step starts on an integer multiple of step size since unix epoch. It's a design decision to work only in UTC which avoids the significant complications of working in a local (and changeable) timezones. I suppose it would be possible to add support for a fixed offset from Unix epoch - but I'm sure if that were added, it wouldn't be long before people started asking to be able to change it. 1267142400 misses that and so RRDtool rounds it until it does fit. > > Once you've done a single update, then that will set the minimum time >> for a future update, and the start time of each consolidated data set >> will be determined by the step time, number of PDPs per CDP, and the >> number of CDPs to keep. >> > >Yes, this should be correct, but I always ad the data at a friday. >If I dump the database, it look like I have stored the data 24h earlier, >this is not nice. You may wish to search the archives, there have been discussions of ways to work around non-zero offsets from UTC for those wanting midnight-midnight (local) which may give you ideas for how to achieve what you want. -- Simon Hobson Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed author Gladys Hobson. Novels - poetry - short stories - ideal as Christmas stocking fillers. Some available as e-books. _______________________________________________ rrd-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
