On Wed, 22 Aug 2012, Jeroen Vermeulen wrote: > On 2012-08-22 10:17, Julian Edwards wrote: > > > It may be possible to coerce OAuth into using a special timestamp > > instead of violently modifying the system clock. It needs a change in > > cloud-init of course. > > Does that violence really hurt though? Who's going to notice our changes > unless we also write the change back to the CMOS clock?
Well, yes, it can hurt. Dann F had explicit experience with it hurting when he was poking around with arm. His particular issue seemed like something we should'nt have hit, but the problem was that there were 2 NICs, the first came up, did dhcp... set the time , that caused havoc on the second that was sleeping. It will just take some playing to see what all falls out when the clock moves forward or backward in a large increment. Luckily, we're pretty early in boot at the point in which this would happen so less things can be affected by it. Also, there is president for this as ntpdate does run on ifup in normal ubuntu systems. I'm not sure why there wasn't fallout seen as a result in other non-maas places. So, in summary: I'm not sure whats going to happen. Reguarding the hwclock, We would *want* to set it IMO. Its better to fix the problem at the source if we can. -- Mailing list: https://launchpad.net/~maas-devel Post to : [email protected] Unsubscribe : https://launchpad.net/~maas-devel More help : https://help.launchpad.net/ListHelp
