On Mon, Aug 23, 2004 at 02:03:22PM -0400, Bush, Andrew G [NY] wrote: > Having searched the archives of both the development and user mailing > lists I am aware that this topic has come up several times. > >From the archives and documentation it is my understanding that RRA > boundaries occur at multiple step sizes from UTC.
Not from UTC (well, that's not the design). They are integer multiples from the unix epoch. > As such whenever trying to general daily summaries the boundaries are > generated at midnight UTC. Not at all. That is a choice, your choice. > Similarly when generating weekly values these seem to occur on a > Wednesday (1/1/1970 was a Thursday) Dito. This just happens if you (let RRDtool) select an RRA with a rowsize of 604800 seconds (7 days). > I am unfortunately working in multiple timezones both on the collection > side and visual display and am measuring discrete items such as fault > calls. > Therefore having the boundaries in the correct place is important. It sure is. That is why you need to carefully design your database. If you need boundaries to occur on -for instance- 05:00 UTC, you have no choice but to keep an RRA with a size of 3600 seconds per interval. That way you can select "--end 05:00 UTC --start end-7d" (do not try this, it is an example for the discussion). "05:00 UTC" translates in "midnight your time". (05:00, or any other appropriate value). If you need to cope with both standard time and daylight saving time, that is just another reason to have 3600 seconds per row. Alex -- I ask you to respect any "Reply-To" and "Mail-Follow-Up" headers. If you reply to me off-list, you'd better tell me you're doing so. If you don't, and if I reply to the list, that's your problem, not mine. -- Unsubscribe mailto:[EMAIL PROTECTED] Help mailto:[EMAIL PROTECTED] Archive http://www.ee.ethz.ch/~slist/rrd-users WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
