On Tue, 2006-08-22 at 13:50 -0400, Steve Magoun wrote: > On Aug 20, 2006, at 12:00 PM, yum-request@xxxxxxxxxxxxxxxxxxxx wrote: > > > I ended up doing something similar in cvs head. > > > > Specifically, I added a check. If the ctime of the cooking file was > > greater than the current time then assume the metadata needs updating. > > > > oh and ntp or rdate is your friend. > > > Thanks for the update on this. A minor nit: shouldn't you check > ctime>=now rather than ctime>now? My assumption is the two values > could only be the same if the clock had been set back at some point > after the cookie was written. > > As I understand it there are reasonably common situations in which > NTP can cause the clock to change by a significant amount, so using > NTP/rdate alone doesn't solve the problem. ntp doesn't move the time more than a few seconds. rdate does - as does ntpdate. but ntp will bail if it has to move it more than a set amount. -sv
