Re: Container Uptime Discussion

Hi Craig,

a nice idea. I wonder if this shouldn't be patched in the  kernel. Just
like reboot(2) has a special behaviour inside a PID namespace perhaps
/proc/uptime should report different values?

Sure, uptime(1) has multiple methods of learning the value and looking
at pid1's start time sounds good, I'd like to know we can get accurate
values from the kernel without resorting to hacks. Hacks may still be
necessary depending on kernel config/security model.

- Kris