Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On 19/01/13 06:36, Neil Bothwick wrote: > On Fri, 18 Jan 2013 18:08:50 + (UTC), Grant Edwards wrote: > By default, ntpd doesn't seem to want to do a step correction to fix large clock errors on startup (there's probably an option for that). >>> >>> That's for ntp-client to do. >> >> In additon to being a server, ntpd _is_ an ntp client. > > But ntpd will not handle large jumps in time, such as when booting with a > broken CMOS clock. So you run /etc/init.d/ntp-client as well, which gets > the clock right at boot time, then ntpd can keep it right. > >> Are you talking about running ntpclient >> (http://doolittle.icarus.com/ntpclient) instead of ntpd? > > No. > > % qlist net-misc/ntp | grep init.d > /etc/init.d/ntpd > /etc/init.d/ntp-client > > It can handle large jumps, check out this in the "man ntp.conf": tinker panic 0 or the -g option to ntpd (again, its in a man page.) Basicly, you just need to read and configure it ... I have it working even on a raspberry pi which has HW clock :) Even the slew rate, or slew/jump can be set. BillK
[gentoo-user] Re: System won't boot if CMOS clock is slow
On 2013-01-18, Neil Bothwick wrote: > On Fri, 18 Jan 2013 18:08:50 + (UTC), Grant Edwards wrote: > >> >> By default, ntpd doesn't seem to want to do >> >> a step correction to fix large clock errors on startup (there's >> >> probably an option for that). >> > >> > That's for ntp-client to do. >> >> In additon to being a server, ntpd _is_ an ntp client. > > But ntpd will not handle large jumps in time, such as when booting > with a broken CMOS clock. So you run /etc/init.d/ntp-client as well, > which gets the clock right at boot time, then ntpd can keep it right. Doh! I couldn't find a package named ntp-client or a program named ntp-client and didn't think to look in /etc/init.d... -- Grant Edwards grant.b.edwardsYow! If Robert Di Niro at assassinates Walter Slezak, gmail.comwill Jodie Foster marry Bonzo??
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Fri, 18 Jan 2013 18:08:50 + (UTC), Grant Edwards wrote: > >> By default, ntpd doesn't seem to want to do > >> a step correction to fix large clock errors on startup (there's > >> probably an option for that). > > > > That's for ntp-client to do. > > In additon to being a server, ntpd _is_ an ntp client. But ntpd will not handle large jumps in time, such as when booting with a broken CMOS clock. So you run /etc/init.d/ntp-client as well, which gets the clock right at boot time, then ntpd can keep it right. > Are you talking about running ntpclient > (http://doolittle.icarus.com/ntpclient) instead of ntpd? No. % qlist net-misc/ntp | grep init.d /etc/init.d/ntpd /etc/init.d/ntp-client -- Neil Bothwick Did you hear about the blind prostitute? You have to hand it to her. signature.asc Description: PGP signature
[gentoo-user] Re: System won't boot if CMOS clock is slow
On 2013-01-17, Neil Bothwick wrote: > On Thu, 17 Jan 2013 22:47:17 + (UTC), Grant Edwards wrote: > >> By default, ntpd doesn't seem to want to do >> a step correction to fix large clock errors on startup (there's >> probably an option for that). > > That's for ntp-client to do. In additon to being a server, ntpd _is_ an ntp client. Are you talking about running ntpclient (http://doolittle.icarus.com/ntpclient) instead of ntpd? -- Grant Edwards grant.b.edwardsYow! I invented skydiving at in 1989! gmail.com
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Thu, 17 Jan 2013 22:47:17 + (UTC), Grant Edwards wrote: > By default, ntpd doesn't seem to want to do > a step correction to fix large clock errors on startup (there's > probably an option for that). That's for ntp-client to do. -- Neil Bothwick Everything should be made as simple as possible, but no simpler. signature.asc Description: PGP signature
[gentoo-user] Re: System won't boot if CMOS clock is slow
On 2013-01-17, Stroller wrote: > > On 16 January 2013, at 16:43, Grant Edwards wrote: > >> I'm having problems with one of my Gentoo systems who's motherboard >> clock is a little slow. When the system comes up, the system time is >> set from the motherboard clock. If that's slow, something in the init >> system seems to panic because some file or other has a timestamp in >> the future. > > You've had lots of other suggestions here, but I think this is > handled fine if you add ntp to the default runlevel (and assuming the > system can connect to the net). It doesn't help problems that occur before ntpd has started and had a chance to slew the clock. By default, ntpd doesn't seem to want to do a step correction to fix large clock errors on startup (there's probably an option for that). FWIW, I recently identified one rather obscure CMOS-clock-related problem scenario (this isn't what happened the other day, but it did waste about half a day a few months back): 1) Your CMOS clock is ahead of the "real" time by several hours for some reason. There are a number of ways this can happen: dual-booting between systems that disagree over UTC vs localtime for the CMOS clock, broken ntpd config, mismanaged timezone settings, etc. 2) Kernel comes up and sets system time from CMOS clock. 3) Root filesystem gets fsck'ed because it's been mounted 28 times, and filesystem meta-data gets timestamp that is actually several hours in the future. 4) System reboots after fsck is finished. 5) Before the recently fsck'ed filesystem gets mounted, the CMOS clock gets reset to the correct time (by dual booting, booting from a rescue CD, or by simply running the BIOS setup and fixing it). 6) The system boots again, and when it tries to mount the root filesystem, the filesystem meta-data has a timestamp that's in the future so the ext3 code in the kernel refuses to mount it. 7) You futz around verifying that you have a good root fs backup, looking at S.M.A.R.T logs and all sorts of other irrelevant things for several hours trying to figure out what's wrong. 8) The universe catches up to the filesystem meta-data timestamp, and suddenly, mysteriously, everything works fine. -- Grant Edwards grant.b.edwardsYow! You mean you don't at want to watch WRESTLING gmail.comfrom ATLANTA?
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Thu, 17 Jan 2013 14:45:13 -0600 Bruce Hill wrote: > On Thu, Jan 17, 2013 at 02:35:50PM -0600, Paul Hartman wrote: > > > > I already had that set. :) I'm not talking about clearing the screen > > at the login prompt; it clears mid-OpenRC during/after the udev step > > (as Grant also described), though in my case it does not halt but > > continues printing the rest of the OpenRC messages and eventually > > gets to the login prompt. I'm thinking it may be fbcondecor > > responsible for the blanking there. I'll need to experiment more > > with it. > > When KMS kicks in? I've never seen KMS clear the screen there for either nvidia or radeon cards so I doubt it's that. All it does at that point that is visible to the user is set the resolution and paint on the $GO_FAST stripes. Now fbcondecor, that's a likely area to look in. I personally no longer use fancy fb and boot splash type things (I like my Alt-F1 console to look like a console and not get in my way) so I doubt I can be of much help further on this. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Thu, Jan 17, 2013 at 02:35:50PM -0600, Paul Hartman wrote: > > I already had that set. :) I'm not talking about clearing the screen > at the login prompt; it clears mid-OpenRC during/after the udev step > (as Grant also described), though in my case it does not halt but > continues printing the rest of the OpenRC messages and eventually gets > to the login prompt. I'm thinking it may be fbcondecor responsible for > the blanking there. I'll need to experiment more with it. When KMS kicks in? -- Happy Penguin Computers >') 126 Fenco Drive ( \ Tupelo, MS 38801 ^^ supp...@happypenguincomputers.com 662-269-2706 662-205-6424 http://happypenguincomputers.com/ Don't top-post: http://en.wikipedia.org/wiki/Top_post#Top-posting
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Wed, Jan 16, 2013 at 2:52 PM, Alan McKinnon wrote: > On Wed, 16 Jan 2013 14:29:04 -0600 > Paul Hartman wrote: > >> On Wed, Jan 16, 2013 at 1:36 PM, Grant Edwards >> wrote: >> > >> > But, in the failures I've been seeing today, it's not getting to >> > agetty. The "clear screen and halt" happens at the "waiting for udev >> > events" step... >> >> FWIW I have also noticed on my machine that somewhere in the middle of >> the OpenRC boot process, the screen gets cleared. Haven't completely >> tracked it down yet. > > In /etc/inittab you need to change stuff like this > > c1:12345:respawn:/sbin/agetty 38400 tty1 linux > > to this > > c1:12345:respawn:/sbin/agetty 38400 tty1 linux --noclear > > > There's a clear elog about this at the end of the util-linux emerge, > perhaps you missed it I already had that set. :) I'm not talking about clearing the screen at the login prompt; it clears mid-OpenRC during/after the udev step (as Grant also described), though in my case it does not halt but continues printing the rest of the OpenRC messages and eventually gets to the login prompt. I'm thinking it may be fbcondecor responsible for the blanking there. I'll need to experiment more with it.
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
> So it is Linux' fault, that your mate used crap Hardware? That is great! > let us blame it for the weather too. And stubbed toes. Well the point was that if OpenBSD had an auto update function I could have installed that and he would still be using OpenBSD happily. If Linux did what OpenBSD does then he would be a happy linux user, well aside from wanting Itunes, though I'm under the impression that's been sorted quite well now. As far as he was concerned he had a fscking watch, what's wrong with this fscking piece of.. or words to that affect and really he was right. The alternative was Vista which took and I mean no joke like 15 mins to finish booting, despite a cleanup and the drive checked out ok. He had just started a gym and couldn't afford extra ram at the time. No need to get touchy, simply real facts, better aired than ignored. Not a great loss or anything. -- ___ 'Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface' (Doug McIlroy) ___
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
So it is Linux' fault, that your mate used crap Hardware? That is great! let us blame it for the weather too. And stubbed toes. Am 16.01.2013 21:43 schrieb "Kevin Chadwick" : > > I have had systems in the past who refused to boot because the > > motherboard time was off, and at first it looked like that was the > > problem again. > > OpenBSD takes the time from the filesystem in that case and boots. I > wish linux did. I had a mate who used to ring me up everytime his mother > in law unplugged the laptop and it was a laptop that's cmos was a pain > to replace. I believe he ended up in 2034 or something after a few > months because I told him the bios key and meant he could avoid > fsck that sometimes gave him various problems =-) > > He was anti slow machines (Vista) and liked linux after being > skeptical. I can't see him trying linux again now :-( > > -- > ___ > > 'Write programs that do one thing and do it well. Write programs to work > together. Write programs to handle text streams, because that is a > universal interface' > > (Doug McIlroy) > ___ > >
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Wed, 16 Jan 2013 14:29:04 -0600 Paul Hartman wrote: > On Wed, Jan 16, 2013 at 1:36 PM, Grant Edwards > wrote: > > > > But, in the failures I've been seeing today, it's not getting to > > agetty. The "clear screen and halt" happens at the "waiting for udev > > events" step... > > FWIW I have also noticed on my machine that somewhere in the middle of > the OpenRC boot process, the screen gets cleared. Haven't completely > tracked it down yet. In /etc/inittab you need to change stuff like this c1:12345:respawn:/sbin/agetty 38400 tty1 linux to this c1:12345:respawn:/sbin/agetty 38400 tty1 linux --noclear There's a clear elog about this at the end of the util-linux emerge, perhaps you missed it > > You can enable logging in your rc.conf to log all of the openrc > output, that might help you see what was on-screen before it got > cleared away. You also might be able to enable interactive mode and go > through it step-by-step to see what's happening in real-time at your > own pace. > -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
> I have had systems in the past who refused to boot because the > motherboard time was off, and at first it looked like that was the > problem again. OpenBSD takes the time from the filesystem in that case and boots. I wish linux did. I had a mate who used to ring me up everytime his mother in law unplugged the laptop and it was a laptop that's cmos was a pain to replace. I believe he ended up in 2034 or something after a few months because I told him the bios key and meant he could avoid fsck that sometimes gave him various problems =-) He was anti slow machines (Vista) and liked linux after being skeptical. I can't see him trying linux again now :-( -- ___ 'Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface' (Doug McIlroy) ___
Re: [gentoo-user] Re: System won't boot if CMOS clock is slow
On Wed, Jan 16, 2013 at 1:36 PM, Grant Edwards wrote: > > But, in the failures I've been seeing today, it's not getting to > agetty. The "clear screen and halt" happens at the "waiting for udev > events" step... FWIW I have also noticed on my machine that somewhere in the middle of the OpenRC boot process, the screen gets cleared. Haven't completely tracked it down yet. You can enable logging in your rc.conf to log all of the openrc output, that might help you see what was on-screen before it got cleared away. You also might be able to enable interactive mode and go through it step-by-step to see what's happening in real-time at your own pace.
[gentoo-user] Re: System won't boot if CMOS clock is slow
On 2013-01-16, Bruce Hill wrote: > On Wed, Jan 16, 2013 at 04:43:16PM +, Grant Edwards wrote: >> I'm having problems with one of my Gentoo systems who's motherboard >> clock is a little slow. When the system comes up, the system time is >> set from the motherboard clock. If that's slow, something in the init >> system seems to panic because some file or other has a timestamp in >> the future. >> >> Just to make it extra convenient, it clears the console screen when >> that happens so there's no actual record of what went wrong or which >> component in th init process is failing. >> >> Going into the BIOS setup and setting the time ahead a minute or two >> will allow the system to start up normally. >> >> Is there any way to disable this "feature"? > > Replace your CMOS battery. I'll try that. I enabled init logging and did more testing, and I now don't think it's actually the motherboard clock that's causing the problem. It seems that a second reboot without changing the clock usually fixes the problem also. I have had systems in the past who refused to boot because the motherboard time was off, and at first it looked like that was the problem again. > Default behavior of agetty is to clear now. In /etc/inittab make sure you have > --noclear in tty1 like this: Yup, figured that one out once I realized that in the normal case it's agetty rather than the init system itself that clears the screen. But, in the failures I've been seeing today, it's not getting to agetty. The "clear screen and halt" happens at the "waiting for udev events" step... -- Grant Edwards grant.b.edwardsYow! You should all JUMP at UP AND DOWN for TWO HOURS gmail.comwhile I decide on a NEW CAREER!!