On Wed, Aug 5, 2009 at 8:12 PM, Bryan J Smith<bjs@xxxxxxxxxx> wrote: > On Thu, 2009-08-06 at 10:53 +0800, John Summerfield wrote: >> I have regularly seen files timestamped "in the future" after >> installation, and I don't do anything extraordinary. I always run my >> computers' clocks on local time. >> I have always assumed, "Stupid Americans, they always think the rest of >> the world does things their way." > > Actually, running the RTC on localtime is very much a legacy Microsoft > limitation / attitude. So the allegedly "stupid American way" is > actually to run your RTC on localtime. > > Open systems have traditionally run their RTC on UTC. The idea is that > you offset everything, system, application, user, etc... localtime, from > UTC, which is the RTC. > > Here's the general rule of thumb ... > - All RTCs and most "real-time" stamps should be made in UTC > - Events should be made in the specific timezone they are local to > > Remember, timezones change, and even timezone rules (offsets) change. > UTC does not. Hence why RTCs should be UTC always, and most real-time > events stamped on UTC. OK. I don't run the RTC on localtime. The vendor/manufacture ships the system configured with the RTC set to some localtime (manufacture site). > > Events, on the other hand, will always be in localtime, so they should > always be stored in localtime with the reference timezone. Again, > because timezones and timezone rules change, offsets should be made from > that stored localtime. E.g., calendar events. Understood. The first time the box get a connection to a NTP server, the RTC gets set to UTC. I'm just trying to find a way to get the RTC close to UTC (within about a minute) without manual intervention during the kickstart prior to package installation if I have a vaild/active network configuration. _______________________________________________ Kickstart-list mailing list Kickstart-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/kickstart-list