On my new Gentoo system, the timezone offset is applied twice at boot. For example:
1. It is 19:00 local time (UTC-5).
2. I reboot.
3. The system time is initially set to 14:00. Later, the time is corrected by the "init.d/hwclock" script.
This initial wrong setting results in warnings such as:
Clock skew detected with `/etc/init.d'
Adjusting mtime of `/run/openrc/deptree' to Sun Dec 17 12:45:57 2017
I followed the Gentoo guides regarding timezone handling. I also asked via IRC and the Forums. Based on that, I think I have configured everything correctly:
$ cat /etc/conf.d/hwclock
# zdump /etc/localtime && date && date -u && /etc/init.d/hwclock show
/etc/localtime Mon Dec 18 22:29:40 2017 -05
Mon Dec 18 22:29:40 -05 2017
Tue Dec 19 03:29:40 UTC 2017
The output of "init.d/hwclock show" matches the UEFI BIOS date and time.
To me, it seems like on boot, when setting the system time, the hwclock time is considered UTC instead of localtime.
Could you link the forum discussions for completeness here?
The bug tracker aims on bugs in ebuilds. Which ebuild(s) (and versions) do you think have a bug?
> Could you link the forum discussions for completeness here?
Here the link: https://forums.gentoo.org/viewtopic-t-1073968.html
> The bug tracker aims on bugs in ebuilds. Which ebuild(s) (and versions) do you
> think have a bug?
I do know which Gentoo ebuild causes this problem. I created the bug in this bug tracker, because the instructions state to do so: "Issues concerning packages in Gentoo and other issues that don't fit in any other component"
As detailed in the forums, I was not able to solve the problem. I had to set my BIOS clock to UTC to get rid of the clock skew warnings at boot.
I am wondering if this could be a bug in OpenRC.
I am not sure how the time setting process works, but I think it is as following:
1. Gentoo boots. The kernel sets the software clock to the value read from the hardware clock. The value is applied as read from the hardware clock, no timezone conversion is applied. So, if your hardware clock is in local time, your software clock is now wrong.
2. Later, "/etc/init.d/hwclock start" is called. This seems to call "settimeofday", which does support timezone conversion. Now, the software clock is correct.
If this is how the process indeed works, then it seems to me that the "settimeofday" should be called before OpenRC checks if the clock if skewed or not.
Assume that if you are the first person to experience this bug, it's probably a local configuration error and not some configuration that should be maintained for all users. We can't fix local problems like yours.
Also, this is not a support forum. You already found two support forums (IRC and the web forums) so there is no reason to inquire about your problem here as well.
I do not agree. I have applied all configurations correctly, as detailed in the forums and also in this bug. If you think this is a configuration problem, then please illustrate here or in the forums which the supposed configuration error is. If there is no configuration problem, then it looks like a bug to me.