1 |
On 18 Jul 2009, at 09:49, Konstantinos Agouros wrote: |
2 |
> ... Is there a |
3 |
> way that the guest recognizes the wakeup and than sets the time |
4 |
> using ntpdate |
5 |
> or based on the clock of the host-os which is ntp-synchronized? |
6 |
|
7 |
I doubt it. I know little about virtualisation, so am ready to be |
8 |
proved wrong on that one. |
9 |
|
10 |
I can only think to make a cron job to call an ntp pool on a regular |
11 |
basis. I think there may be a switch to ntp which allows it to make |
12 |
larger jumps. |
13 |
|
14 |
The problem with this is that etiquette would tend to dictate not |
15 |
syncing with the upstream ntp server more than once per hour, so each |
16 |
time you wake it up your clock is going to be wrong for an average of |
17 |
45 minutes (allowing for Sod's Law). Solution that occurs to me is to |
18 |
run your own local ntp server on another machine which doesn't sleep |
19 |
(or on the host Mac itself) and sync with that; thus your cron job can |
20 |
run every minute to update the clock, and the time it takes to correct |
21 |
after waking will be insignificant. |
22 |
|
23 |
Or sleep or hibernate the guest o/s before sleeping the host. |
24 |
|
25 |
Stroller. |
26 |
|
27 |
|
28 |
|
29 |
PS: shouldn't the correct virtualisation terminology be "host o/s" and |
30 |
"parasite o/s"? Methinks the marketing department were allowed to |
31 |
write the docs on this one. |