Bug 1792652
Summary: | boots to GMT/Zulu time until network on | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Nick Levinson <Nick_Levinson> |
Component: | systemd | Assignee: | systemd-maint |
Status: | CLOSED CANTFIX | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | medium | Docs Contact: | |
Priority: | unspecified | ||
Version: | 31 | CC: | lnykryn, msekleta, ssahani, s, systemd-maint, zbyszek |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | x86_64 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2020-03-31 14:17:17 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Nick Levinson
2020-01-18 21:17:36 UTC
It sounds like your RTC battery died. I think the best solution would be to simply exchange the battery (it can be removed and replaced w/o any special tools usually). On the software side, it sounds like you machine sets the RTC clock to local time. This is a generally broken concept that cannot be made to work. As a work-around, you can tell the system that RTC is using localtime with 'timedatectl set-local-rtc 1'. If you have other OSes installed on the same machine, this should work OK. You're likely right that the CMOS battery should be replaced, but I use it daily but cold-boot only every few weeks or couple of months or so and the BIOS password still works. Since the error is not in the time displayed by BIOS after I set it but is in what is displayed by the OS/DE even later, wrong by just a few time zones, I doubt that's a battery problem. With info timedatectl I see timedatectl status --all has RTC=UTC so I think the easiest way out for me is to live with the error and let online snchronization take care of it albeit ith a bit of a delay. It has only one OS, with a choice of kernels, unless I boot from a live CD. I'm leaving the bug status as is. Thanks. I didn't know about timedatectl. > timedatectl status --all has RTC=UTC
Yes, and 'timedatectl set-local-rtc 1' would switch it to "local" mode.
OK, I think that there's no bug here, in the sense that there's a known failure
stemming from BIOS and Microsoft Windows insisting to use RTC in local mode that
we can't do much about.
That's puzzling, partly since I was quite thorough in erasing everything from the HDD (running DBAN) and see no trace of Windows. But if the Dell laptop's BIOS was designed to support Windows in a way that Linux can't access and correct for, that would be interesting; or maybe timedatectl is intended to correct for it. info timedatectl describes a limitation on set-local-rtc 1 that makes it more trouble than it's worth unless I maintain an offline note to myself to manually update between daylight and standard and woe to frequent travellers crossing zones. Yesterday, I was online with the correct local time (within a minute or two) but when I lost my Internet connection it reverted to GMT until I got a network connection again. Also, both times yesterday, auto-coordinating the time to local time took longer than usual. So maybe you're right that it's likely a BIOS battery problem. The coin cell battery module was about 11 years old, so it was about time to replace it. I replaced the coin cell battery (CR2032) module with a new OEM one, and I haven't seen a time zone error since. I don't even have to enter the current date and time anymore. When I swap the main battery without AC power, I usually complete the swap within just a few seconds, and maybe if I took longer Setup would forget the date, time, and zone. I have not tested a longer interruption between swaps. Also, since replacing the module, I have not seen a problem after going offline about the desktop showing non-local GMT. So, this seems to confirm the diagnosis in comment 1, not quite that the battery died but that it was dying. Thank you. On the replacement procedure: http://brittlebit.org/hardware/replacing-the-coin-cell-battery-in-the-dell-latitude-e6400.html |