[opensuse] 10.2 system clock too fast and NTP
Hi, I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast. This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue. -- Anders Norrbring Norrbring Consulting
Anders Norrbring wrote:
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue.
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted... -- "In theory, there is no difference between theory and practice. But, in practice, there is." - Jan L.A. Van De Snepscheut -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Jan Karjalainen a écrit :
Anders Norrbring wrote:
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue.
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
ntpd is made to cope with this, but I suspect there is a limit to the fix it can do :-) jdd -- http://www.dodin.net http://dodin.org/mediawiki/index.php/GPS_Lowrance_GO -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 00:37, Jan Karjalainen wrote:
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
That's silly. Just run the daemon, that's what its for. man ntpd -- _____________________________________ John Andersen
John Andersen wrote:
On Saturday 16 December 2006 00:37, Jan Karjalainen wrote:
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
That's silly. Just run the daemon, that's what its for.
man ntpd
Silly is that if I let ntp run by itself, the clock will be off by hours after only one day... -- "In theory, there is no difference between theory and practice. But, in practice, there is." - Jan L.A. Van De Snepscheut -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 01:03, Jan Karjalainen wrote:
John Andersen wrote:
On Saturday 16 December 2006 00:37, Jan Karjalainen wrote:
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
That's silly. Just run the daemon, that's what its for.
man ntpd
Silly is that if I let ntp run by itself, the clock will be off by hours after only one day...
I think you are confusing the command line ntp with the always running and always correcting ntpd. You have to configure ntpd by adding server lines in /etc/ntp.conf but once you do that if your clock is close at boot time it will keep it in sync forever. But perhaps the bets solution is to find out why your clock runs that fast. -- _____________________________________ John Andersen
I think you are confusing the command line ntp with the always running and always correcting ntpd.
You have to configure ntpd by adding server lines in /etc/ntp.conf but once you do that if your clock is close at boot time it will keep it in sync forever.
But perhaps the bets solution is to find out why your clock runs that fast.
I have the ntp daemon configured with a working server, and it produces no error logs. It´s just that the clock is really running way too fast for the ntp daemon to work. If I let the ntp daemon run as it is supposed to, the clock will be off by hours after one day. Adding "rcntp restart" every 10 minutes to crontab helps, but does not solve the problem with the fast clock. -- "In theory, there is no difference between theory and practice. But, in practice, there is." - Jan L.A. Van De Snepscheut -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Jan Karjalainen wrote:
I think you are confusing the command line ntp with the always running and always correcting ntpd.
You have to configure ntpd by adding server lines in /etc/ntp.conf but once you do that if your clock is close at boot time it will keep it in sync forever.
But perhaps the bets solution is to find out why your clock runs that fast.
I have the ntp daemon configured with a working server, and it produces no error logs. It´s just that the clock is really running way too fast for the ntp daemon to work. If I let the ntp daemon run as it is supposed to, the clock will be off by hours after one day. Adding "rcntp restart" every 10 minutes to crontab helps, but does not solve the problem with the fast clock.
It only helps because the daemon will set the time at start. You might as well execute ntpdate in a cronjob. Is this an installation running within a VM? I have the same phenomenon with one of my vmware installations. xntpd doesn't work on that system either. Sandy -- List replies only please! Please address PMs to: news-reply2 (@) japantest (.) homelinux (.) com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
John Andersen wrote:
On Saturday 16 December 2006 02:32, Sandy Drobic wrote:
It only helps because the daemon will set the time at start. You might as well execute ntpdate in a cronjob.
Correction: The daemon will set the time continuously.
In theory. (^-^) On most of my system it works flawlessly. Just on this one, nothing happens. For a test I even set logconfig = all logfile = /var/log/ntp /var/log/ntp: 16 Dec 13:09:26 ntpd[3915]: system event 'event_restart' (0x01) status 'sync_alarm, sync_unspec, 1 event, event_unspec' (0xc010) There are no sync events recorded aside at the start of the xntpd. Since I am only using it to test the update from Suse 9.2 to Suse 10.2 I won't debug this any further. I don't see this with a 10.2 kernel. Sandy -- List replies only please! Please address PMs to: news-reply2 (@) japantest (.) homelinux (.) com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Sandy Drobic wrote:
Jan Karjalainen wrote:
I think you are confusing the command line ntp with the always running and always correcting ntpd.
You have to configure ntpd by adding server lines in /etc/ntp.conf but once you do that if your clock is close at boot time it will keep it in sync forever.
But perhaps the bets solution is to find out why your clock runs that fast. I have the ntp daemon configured with a working server, and it produces no error logs. It´s just that the clock is really running way too fast for the ntp daemon to work. If I let the ntp daemon run as it is supposed to, the clock will be off by hours after one day. Adding "rcntp restart" every 10 minutes to crontab helps, but does not solve the problem with the fast clock.
It only helps because the daemon will set the time at start. You might as well execute ntpdate in a cronjob.
Is this an installation running within a VM? I have the same phenomenon with one of my vmware installations. xntpd doesn't work on that system either.
Sandy No, it´s not a VM.
-- "In theory, there is no difference between theory and practice. But, in practice, there is." - Jan L.A. Van De Snepscheut -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
John Andersen wrote:
On Saturday 16 December 2006 00:37, Jan Karjalainen wrote:
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
That's silly. Just run the daemon, that's what its for.
man ntpd
Thats not silly, running ntpd every 10 minutes is silly. Clearly there is a problem with his machine. I don't think NTPD is really intended for this situation. Mark -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 01:30, Mark Hounschell wrote:
John Andersen wrote:
On Saturday 16 December 2006 00:37, Jan Karjalainen wrote:
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
That's silly. Just run the daemon, that's what its for.
man ntpd
Thats not silly, running ntpd every 10 minutes is silly. Clearly there is a problem with his machine. I don't think NTPD is really intended for this situation.
Mark
Sigh... Nobody suggested he run nptd every ten minutes. Its a daemon. (that's why it has a d on the end of the name). It runs continuously. Its intended for PRECISELY this sort of thing. -- _____________________________________ John Andersen
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2006-12-16 at 01:37 -0900, John Andersen wrote:
I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
That's silly. Just run the daemon, that's what its for.
man ntpd
Thats not silly, running ntpd every 10 minutes is silly. Clearly there is a problem with his machine. I don't think NTPD is really intended for this situation.
Sigh...
Nobody suggested he run nptd every ten minutes. Its a daemon. (that's why it has a d on the end of the name). It runs continuously.
Its intended for PRECISELY this sort of thing.
But within limits. After setting the clock once, the clock is running so fast that ntpd can't cope. It will try to slew the clock back, but it does so slowly. Soon the time will be out of range, and ntpd will stop trying because it has already gone beyond its sanity limit. Thus the user restarts it. "rcntp restart" works in this situation because it sets the clock at the start even if the jump is big: no sanity checks during service start. It would be better to use "rcntp ntptimeset" as a cron job every 5 or 10 minutes in this situation as a hack till the real problem can be found. Something in the kernel, I suppose. It should go to bugzilla, I think. The trick I often say of removing the "/etc/adjtime" will not work this time, because the problem arises after the system is running and having adjusted the time properly, I understand. Try erasing ntpd adjustments, not sure exactly which ATM. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFg/5ZtTMYHG2NR9URAtzFAJ9xEaHNfcWviDluKOjFTts9HkdufwCeIRYA E3ItxctNJ61Itl53CASFhZ8= =aO1W -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 05:10, Carlos E. R. wrote:
But within limits. After setting the clock once, the clock is running so fast that ntpd can't cope. It will try to slew the clock back, but it does so slowly.
That too is configurable. You can tell it not to try these slow movements, but do it in one big step. While not ideal, this is no worse for the system than running ntpdate out of cron. -- _____________________________________ John Andersen
On Saturday 16 December 2006 10:37, Jan Karjalainen wrote:
Anders Norrbring wrote:
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue.
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
In my experiences ntp deamon adjust the time only if the difference is less than 3600 seconds. You can see in /var/log/ntp if there is a message like I had it: "time correction of -3600 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time." My inital "time-problem" with a system clock running way too fast was discussed here: http://lists.suse.com/archive/suse-linux-e/2006-Mar/2932.html The suggested and helpful solution by Carlos E.R. in http://lists.suse.com/archive/suse-linux-e/2006-Mar/3273.html was: setup the clock by your prefered method hwclock --systohc rm /etc/adjtime It's important to remove /etc/adjtime after adjusting the time manually. The system will create a new /etc/adjtime later. Since I did the above my clock is as perfect as can be. regards Danie -- Daniel Bauer photographer Basel Switzerland professional photography: http://www.daniel-bauer.com Madagascar special: http://www.sanic.ch -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 02:42, Daniel Bauer wrote:
In my experiences ntp deamon adjust the time only if the difference is less than 3600 seconds. You can see in /var/log/ntp if there is a message like I had it:
"time correction of -3600 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time."
That's what command line switch -g is for. -- _____________________________________ John Andersen
John Andersen a écrit :
On Saturday 16 December 2006 02:42, Daniel Bauer wrote:
In my experiences ntp deamon adjust the time only if the difference is less than 3600 seconds. You can see in /var/log/ntp if there is a message like I had it:
"time correction of -3600 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time."
That's what command line switch -g is for.
RTFM: Most operating systems and hardware of today incorporate a time-of-year (TOY) chip to maintain the time during periods when the power is off. When the machine is booted, the chip is used to initialize the operating system time. After the machine has synchronized to a NTP server, the operating system corrects the chip from time to time. In case there is no TOY chip or for some reason its time is more than 1000s from the server time, ntpd assumes something must be terribly wrong and the only reliable action is for the operator to intervene and set the clock by hand. This causes ntpd to exit with a panic message to the system log. The -g option overrides this check and the clock will be set to the server time regardless of the chip time. However, and to protect against broken hardware, such as when the CMOS battery fails or the clock counter becomes defective, once the clock has been set, an error greater than 1000s will cause ntpd to exit anyway. jdd -- http://www.dodin.net http://dodin.org/mediawiki/index.php/GPS_Lowrance_GO -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
jdd wrote:
John Andersen a écrit :
On Saturday 16 December 2006 02:42, Daniel Bauer wrote:
In my experiences ntp deamon adjust the time only if the difference is less than 3600 seconds. You can see in /var/log/ntp if there is a message like I had it:
"time correction of -3600 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time."
That's what command line switch -g is for.
RTFM:
Most operating systems and hardware of today incorporate a time-of-year (TOY) chip to maintain the time during periods when the power is off. When the machine is booted, the chip is used to initialize the operating system time. After the machine has synchronized to a NTP server, the operating system corrects the chip from time to time. In case there is no TOY chip or for some reason its time is more than 1000s from the server time, ntpd assumes something must be terribly wrong and the only reliable action is for the operator to intervene and set the clock by hand. This causes ntpd to exit with a panic message to the system log. The -g option overrides this check and the clock will be set to the server time regardless of the chip time. However, and to protect against broken hardware, such as when the CMOS battery fails or the clock counter becomes defective, once the clock has been set, an error greater than 1000s will cause ntpd to exit anyway.
jdd
Dude, my hardware is not broke! OpenSUSE 10.2 is broke! Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Anders Norrbring wrote:
ByteEnable wrote: [8<]
Dude, my hardware is not broke! OpenSUSE 10.2 is broke!
Byte
Watch the attitude... If you don't like it the way it is, either rewrite openSUSE, or change distro, it's your choice.
I think you need to check yours little boy. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 23:53, ByteEnable wrote:
Anders Norrbring wrote:
ByteEnable wrote: [8<]
Dude, my hardware is not broke! OpenSUSE 10.2 is broke!
Byte
Watch the attitude... If you don't like it the way it is, either rewrite openSUSE, or change distro, it's your choice.
I think you need to check yours little boy.
keep your haiy knickers on boy you have been asked what type of CPU you have you have NOT furnished an Answer leat not by 07:01 Sunday 17th 12 2006 you have not so furnish a few sensible answers and less with the sacasam you may get somewhere .. Pete . -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
ByteEnable a écrit :
Dude, my hardware is not broke! OpenSUSE 10.2 is broke!
or the particular ntp version of 10.2... jdd -- http://www.dodin.net http://dodin.org/mediawiki/index.php/GPS_Lowrance_GO -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2006-12-16 at 16:38 +0100, jdd wrote:
Dude, my hardware is not broke! OpenSUSE 10.2 is broke!
or the particular ntp version of 10.2...
That problem of the clock going to fast has appeared and dissapeared randomly over the versions, but I have no idea why or what could be the cause. Some change in the kernel, I guess. I don't think it is ntpd, but rather that ntpd is unable to compensate the problem - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFhBdLtTMYHG2NR9URAttQAJ95wHE73aPC0Go7A39BPQaeYruWLwCfaJqa TiDiU806A380gmrJACRavzg= =shxk -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Saturday 2006-12-16 at 16:38 +0100, jdd wrote:
Dude, my hardware is not broke! OpenSUSE 10.2 is broke!
Well in this entire thread, unless I missed it, did he ever really say he DID remove /etc/adjtime properly? I saw a "won't work"...but did he
On Sat, 2006-12-16 at 16:56 +0100, Carlos E. R. wrote: try? If he removes it, manually sets the time and removes it once again, will the pc keep reasonably close time on it's own? ie WITHOUT ntpd? Tom -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2006-12-16 at 09:56 -0700, Tom Patton wrote:
Well in this entire thread, unless I missed it, did he ever really say he DID remove /etc/adjtime properly? I saw a "won't work"...but did he try?
My crystal ball says it won't. O:-)
If he removes it, manually sets the time and removes it once again, will the pc keep reasonably close time on it's own? ie WITHOUT ntpd?
I don't think so. Not in this case. That trick only works when the time is set wrong after booting, but it doesn't when the clock goes wrong after that precise moment. There is another trick, however, to correct the system time without ntpd, using the cmos clock periodically for the update. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFhC5ttTMYHG2NR9URAiiaAJ9h0x0v0Zp5DQJskCyHpA2/Nc/FeQCgifDO nbW6qubgQjd3qk6XR2F1zDE= =lP77 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Saturday 2006-12-16 at 09:56 -0700, Tom Patton wrote:
Well in this entire thread, unless I missed it, did he ever really say he DID remove /etc/adjtime properly? I saw a "won't work"...but did he try?
My crystal ball says it won't. O:-)
If he removes it, manually sets the time and removes it once again, will the pc keep reasonably close time on it's own? ie WITHOUT ntpd?
I don't think so. Not in this case. That trick only works when the time is set wrong after booting, but it doesn't when the clock goes wrong after that precise moment.
There is another trick, however, to correct the system time without ntpd, using the cmos clock periodically for the update.
And also, as I suggested initially, set the boot parameter 'clock=pit'... -- Anders Norrbring Norrbring Consulting
Daniel Bauer wrote:
On Saturday 16 December 2006 10:37, Jan Karjalainen wrote:
Anders Norrbring wrote:
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue.
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
In my experiences ntp deamon adjust the time only if the difference is less than 3600 seconds. You can see in /var/log/ntp if there is a message like I had it:
"time correction of -3600 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time."
My inital "time-problem" with a system clock running way too fast was discussed here: http://lists.suse.com/archive/suse-linux-e/2006-Mar/2932.html
The suggested and helpful solution by Carlos E.R. in http://lists.suse.com/archive/suse-linux-e/2006-Mar/3273.html was:
setup the clock by your prefered method hwclock --systohc rm /etc/adjtime
It's important to remove /etc/adjtime after adjusting the time manually. The system will create a new /etc/adjtime later.
Since I did the above my clock is as perfect as can be.
regards
Danie
I've tried what you have stated above. It does not help. After five hours the clock was still off by -1060.096981 seconds according to ntpdate time-a.nist.gov. I've tried NTP daemon too: 14 Dec 22:26:39 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:26:37 ntpd[12811]: time reset -1.472229 s 14 Dec 22:26:37 ntpd[12811]: kernel time sync enabled 0001 14 Dec 22:28:18 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:29:07 ntpd[12811]: no servers reachable 14 Dec 22:29:45 ntpd[12811]: synchronized to 129.6.15.29, stratum 1 14 Dec 22:30:24 ntpd[12811]: no servers reachable 14 Dec 22:30:33 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:31:15 ntpd[12811]: synchronized to LOCAL(0), stratum 10 14 Dec 22:32:56 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:33:59 ntpd[12811]: synchronized to LOCAL(0), stratum 10 14 Dec 22:35:15 ntpd[12811]: ntpd exiting on signal 15 Bottom line is that something in OpenSUSE 10.2 is messed up! Too many people complaining about the same issue. Again, I've only had this issue with OpenSUSE 10.2, other Linux distro's have worked flawlessly. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
ByteEnable wrote:
Daniel Bauer wrote:
On Saturday 16 December 2006 10:37, Jan Karjalainen wrote:
Anders Norrbring wrote:
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue.
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
In my experiences ntp deamon adjust the time only if the difference is less than 3600 seconds. You can see in /var/log/ntp if there is a message like I had it:
"time correction of -3600 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time."
My inital "time-problem" with a system clock running way too fast was discussed here: http://lists.suse.com/archive/suse-linux-e/2006-Mar/2932.html
The suggested and helpful solution by Carlos E.R. in http://lists.suse.com/archive/suse-linux-e/2006-Mar/3273.html was:
setup the clock by your prefered method hwclock --systohc rm /etc/adjtime
It's important to remove /etc/adjtime after adjusting the time manually. The system will create a new /etc/adjtime later.
Since I did the above my clock is as perfect as can be.
regards
Danie
I've tried what you have stated above. It does not help. After five hours the clock was still off by -1060.096981 seconds according to ntpdate time-a.nist.gov.
I've tried NTP daemon too:
14 Dec 22:26:39 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:26:37 ntpd[12811]: time reset -1.472229 s 14 Dec 22:26:37 ntpd[12811]: kernel time sync enabled 0001 14 Dec 22:28:18 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:29:07 ntpd[12811]: no servers reachable 14 Dec 22:29:45 ntpd[12811]: synchronized to 129.6.15.29, stratum 1 14 Dec 22:30:24 ntpd[12811]: no servers reachable 14 Dec 22:30:33 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:31:15 ntpd[12811]: synchronized to LOCAL(0), stratum 10 14 Dec 22:32:56 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:33:59 ntpd[12811]: synchronized to LOCAL(0), stratum 10 14 Dec 22:35:15 ntpd[12811]: ntpd exiting on signal 15
Bottom line is that something in OpenSUSE 10.2 is messed up! Too many people complaining about the same issue.
Again, I've only had this issue with OpenSUSE 10.2, other Linux distro's have worked flawlessly.
Byte
You're probably wrong. I run 18 servers on 10.2, both physical and virtual under VMware Server and VMware ESX, and I don't have any problems whatsoever with neither the clock nor the ntpd subsystem. Bottom line; SUSE works fine, but on your particular system, it doesn't. -- Anders Norrbring Norrbring Consulting
On Sat December 16 2006 10:21 am, ByteEnable scratched these words onto a coconut shell, hoping for an answer: <Snip>
Bottom line is that something in OpenSUSE 10.2 is messed up! Too many people complaining about the same issue.
Again, I've only had this issue with OpenSUSE 10.2, other Linux distro's have worked flawlessly.
Gee, two out of , how many 10k, more? , people have this problem.... ? Suse 10.2 must be really messed up. <sigh> I suspect the guys are right and there is simply something about your setup that doesn't work as expected. If other distros work perfectly.. well go there , wait for a Suse 10.2 update , or whatever.. it wont help your problem to get this group mad at you, or discount what you say, now will it? -- j -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2006-12-16 at 12:23 -0500, jfweber@gilweber.com wrote:
On Sat December 16 2006 10:21 am, ByteEnable scratched these words onto a coconut shell, hoping for an answer: <Snip>
Bottom line is that something in OpenSUSE 10.2 is messed up! Too many people complaining about the same issue.
Again, I've only had this issue with OpenSUSE 10.2, other Linux distro's have worked flawlessly.
Gee, two out of , how many 10k, more? , people have this problem.... ? Suse 10.2 must be really messed up. <sigh>
I suspect the guys are right and there is simply something about your setup that doesn't work as expected. If other distros work perfectly.. well go there , wait for a Suse 10.2 update , or whatever.. it wont help your problem to get this group mad at you, or discount what you say, now will it?
Don't discount him so easily. He must be very frustrated now if it worked before and doesn't now. I have heard of this very same problem on previous SuSE versions, it is not new, and the cause is unknown - for me at least. Try to understand his situation :-) - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFhC8qtTMYHG2NR9URAt+ZAJ9o45PYh1P7wjcroePFzjXO3l0w9gCeMGyr OKsvCpUYGHUGNlLETYUDWxs= =X4/3 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
jfweber@gilweber.com wrote:
On Sat December 16 2006 10:21 am, ByteEnable scratched these words onto a coconut shell, hoping for an answer: <Snip>
Bottom line is that something in OpenSUSE 10.2 is messed up! Too many people complaining about the same issue.
Again, I've only had this issue with OpenSUSE 10.2, other Linux distro's have worked flawlessly.
Gee, two out of , how many 10k, more? , people have this problem.... ? Suse 10.2 must be really messed up. <sigh>
I suspect the guys are right and there is simply something about your setup that doesn't work as expected. If other distros work perfectly.. well go there , wait for a Suse 10.2 update , or whatever.. it wont help your problem to get this group mad at you, or discount what you say, now will it?
Why don't you offer suggestions instead of little boy threats? Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
* ByteEnable
Why don't you offer suggestions instead of little boy threats?
Congradulations, you have attained _troll_ status <plonk> -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 14:55, ByteEnable wrote:
Why don't you offer suggestions instead of little boy threats?
Byte
I'm pretty sure JFWeber doesn't make Little Boy threats. Unless, of course, YOU feel threatened by her. -- _____________________________________ John Andersen -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
jfweber@gilweber.com wrote:
On Sat December 16 2006 10:21 am, ByteEnable scratched these words onto a coconut shell, hoping for an answer: <Snip>
Bottom line is that something in OpenSUSE 10.2 is messed up! Too many people complaining about the same issue.
Again, I've only had this issue with OpenSUSE 10.2, other Linux distro's have worked flawlessly.
Gee, two out of , how many 10k, more? , people have this problem.... ? Suse 10.2 must be really messed up. <sigh>
I suspect the guys are right and there is simply something about your setup that doesn't work as expected. If other distros work perfectly.. well go there , wait for a Suse 10.2 update , or whatever.. it wont help your problem to get this group mad at you, or discount what you say, now will it?
Why don't you offer suggestions instead of little boy threats? I'm not a little *boy* Geez, you male centric git! I offered a suggestion as valid as some others.. if one can't reprogram stuff one must wait until those who can, do.. I can't , so I mention
On Sat December 16 2006 6:55 pm, ByteEnable scratched these words onto a coconut shell, hoping for an answer: problem, and when I'm told it's a bug, I wait til it's fixed. Whining doesn't help little boy... Nanny gets very cranky and you'll get no tea! ;) -- j Merry Frikin' Winter.. I wish you all Boat Drinks, but I'm heading South before my brain shrinks. (w/ appologies to Mr. Buffett) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 06:21, ByteEnable wrote:
4 Dec 22:26:39 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:26:37 ntpd[12811]: time reset -1.472229 s 14 Dec 22:26:37 ntpd[12811]: kernel time sync enabled 0001 14 Dec 22:28:18 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:29:07 ntpd[12811]: no servers reachable 14 Dec 22:29:45 ntpd[12811]: synchronized to 129.6.15.29, stratum 1 14 Dec 22:30:24 ntpd[12811]: no servers reachable 14 Dec 22:30:33 ntpd[12811]: synchronized to 128.138.140.44, stratum 1
I'd be pretty suspicious of those "no servers reachable" messages. Do you have one or several servers configured? You should use as many as 4 or 5, and <dons flamesuit> avoid the pools because they have proven to be a single point of failure IMHO. If you use a pool, make sure you have at least two other clocks named explicitly. Universities (at least many of them in the US) seem to have time servers and mega-bandwidth. I use their servers a lot. -- _____________________________________ John Andersen
On Saturday 16 December 2006 23:03, John Andersen wrote:
Do you have one or several servers configured? You should use as many as 4 or 5, and <dons flamesuit> avoid the pools because they have proven to be a single point of failure IMHO. If you use a pool, make sure you have at least two other clocks named explicitly.
How about the option "Use Random Servers from pool.ntp.org" in openSUSE 10.2? YaST puts {0,1,2}.pool.ntp.org in /etc/ntp.conf, of which the DNS-records change every hour. Or do you still stand by your advice I quoted? Cheers, Leen -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 13:20, Leendert Meyer wrote:
On Saturday 16 December 2006 23:03, John Andersen wrote:
Do you have one or several servers configured? You should use as many as 4 or 5, and <dons flamesuit> avoid the pools because they have proven to be a single point of failure IMHO. If you use a pool, make sure you have at least two other clocks named explicitly.
How about the option "Use Random Servers from pool.ntp.org" in openSUSE 10.2? YaST puts {0,1,2}.pool.ntp.org in /etc/ntp.conf, of which the DNS-records change every hour.
Or do you still stand by your advice I quoted?
That was exactly what my advice was referring to. For some reason that has never worked for me in a reliable way. Perhaps occasionally that url is unreachable or something. I prefer selecting a server that I have known to be workable and reliable. -- _____________________________________ John Andersen
On 2006-12-16 16:03, John Andersen wrote:
On Saturday 16 December 2006 06:21, ByteEnable wrote:
4 Dec 22:26:39 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:26:37 ntpd[12811]: time reset -1.472229 s 14 Dec 22:26:37 ntpd[12811]: kernel time sync enabled 0001 14 Dec 22:28:18 ntpd[12811]: synchronized to 128.138.140.44, stratum 1 14 Dec 22:29:07 ntpd[12811]: no servers reachable 14 Dec 22:29:45 ntpd[12811]: synchronized to 129.6.15.29, stratum 1 14 Dec 22:30:24 ntpd[12811]: no servers reachable 14 Dec 22:30:33 ntpd[12811]: synchronized to 128.138.140.44, stratum 1
I'd be pretty suspicious of those "no servers reachable" messages.
Do you have one or several servers configured? You should use as many as 4 or 5, and <dons flamesuit> avoid the pools because they have proven to be a single point of failure IMHO. If you use a pool, make sure you have at least two other clocks named explicitly.
Universities (at least many of them in the US) seem to have time servers and mega-bandwidth. I use their servers a lot.
Both those IPs resolve; 128.138.140.44 is UColorado, while 129.6.15.29 is time-b.nist.gov. AFAIK, neither of those is in any ntp pool. The "no servers reachable" messages don't bother me very much, as ntpd should be able to keep the time within reason for a lot longer than the interval in the above log fragment. What is of more concern is that the local clock seems way out of touch with reality. That suggests to me that ntp is working properly, so long as it can reach a server of course. It is rather something in the kernel configuration that is wrong. For Byte: what is the output of "adjtimex -p"? -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2006-12-16 at 16:35 -0600, Darryl Gregorash wrote:
Both those IPs resolve; 128.138.140.44 is UColorado, while 129.6.15.29 is time-b.nist.gov. AFAIK, neither of those is in any ntp pool.
The "no servers reachable" messages don't bother me very much, as ntpd should be able to keep the time within reason for a lot longer than the interval in the above log fragment.
And that is a temporary condition. However, it is good to list several different servers at different locations. I use servers from different pools from different countries: more than a dozen. If one is not available or is bad, the daemon will use another one. Actually, it will choose the best one from the list. It also occurs to me that the OP having selected a stratum 1 server, he could be rejected service sometimes.
What is of more concern is that the local clock seems way out of touch with reality. That suggests to me that ntp is working properly, so long as it can reach a server of course. It is rather something in the kernel configuration that is wrong.
Probably.
For Byte: what is the output of "adjtimex -p"?
That would be very interesting, good idea. It could perhaps be used later to keep the clock on his toes. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFhI0qtTMYHG2NR9URAhFuAJ4o7BnTCt2m3mU2O2TiBASOTIey7QCfQgR3 sSezwOe1KlaDFFqVqG44UeA= =GyLG -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
<snip> It also occurs to me that the OP having selected a stratum 1 server, he could be rejected service sometimes. I am pretty sure the two servers in his log fragment are both open --
On 2006-12-16 18:19, Carlos E. R. wrote: the one at NIST certainly is; I have used it in the past.
What is of more concern is that the local clock seems way out of touch with reality. That suggests to me that ntp is working properly, so long as it can reach a server of course. It is rather something in the kernel configuration that is wrong.
Probably.
For Byte: what is the output of "adjtimex -p"?
That would be very interesting, good idea. It could perhaps be used later to keep the clock on his toes.
I am not thinking of using adjtimex to adjust the clock, and indeed it should not be so used when ntpd is running. However, with the -p parameter, adjtimex prints the kernel time parameters, including the value of "tick". If there is an easily resolved problem here, the value of "tick" is the first place to look. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2006-12-16 at 18:38 -0600, Darryl Gregorash wrote:
It also occurs to me that the OP having selected a stratum 1 server, he could be rejected service sometimes. I am pretty sure the two servers in his log fragment are both open -- the one at NIST certainly is; I have used it in the past.
Ok, but it is a fact that he gets the "no servers reachable" message now and then. I wouldn't hurt to define more servers.
For Byte: what is the output of "adjtimex -p"?
That would be very interesting, good idea. It could perhaps be used later to keep the clock on his toes. I am not thinking of using adjtimex to adjust the clock, and indeed it should not be so used when ntpd is running.
No simultaneously, of course. I said "later". Check what the command says, and perhaps, use it manually to slow the system clock to a reasonable value.
However, with the -p parameter, adjtimex prints the kernel time parameters, including the value of "tick". If there is an easily resolved problem here, the value of "tick" is the first place to look.
I know. He also could try erasing /var/lib/ntp/drift/ntp.drift first before starting ntpd again. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFhJ9TtTMYHG2NR9URAuplAJ9+UDar4TxQ8SksvYVCXPFQg7dafgCfY1eu z2VbOEaLMzTHI9wuG+/rbHY= =bBQ8 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Jan Karjalainen wrote:
Anders Norrbring wrote:
ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Are you possibly running this in a VMware virtual machine? If you are, add 'clock=pit' into your grub boot parameters. It's a known VMware issue.
I have the same problem with one of my machines, the clock runs way too fast. I have to run "rcntp restart" every 10 minutes to keep it somehow adjusted...
Are you running with the configuration from YAST? Ntp is suppose to have some files that are used to adjust the clock speed. When you run ntp regularly, it will adjust these files automatically and the kernel will use these files to set it up properly. By the way, you can adjust the amount of time that it will look at the ntp server. It is in the ntp configuration file. -- Joseph Loo jloo@acm.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2006-12-16 at 02:40 -0600, ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Byte
On Sat, 2006-12-16 at 02:40 -0600, ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
Byte
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2006-12-16 at 21:09 -0600, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Byte
Then I have to ask again, have you deleted the /etc/adjtime file? My understanding is that sets the correction factor for the kernel "system clock". Get it to run correctly before you start messing with NTPD, IMO. Tom -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2006-12-16 at 22:13 -0700, Tom Patton wrote:
On Sat, 2006-12-16 at 21:09 -0600, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Byte
Then I have to ask again, have you deleted the /etc/adjtime file? My understanding is that sets the correction factor for the kernel "system clock". Get it to run correctly before you start messing with NTPD, IMO.
Tom
Yeah, I deleted the /etc/adjtime file, played with hwclock, played with ntp, etc. Its just no workie. <sigh>. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2006-12-17 00:00, ByteEnable wrote:
<snip> Yeah, I deleted the /etc/adjtime file, played with hwclock, played with ntp, etc. Its just no workie. <sigh>.
Byte
Stop ntpd, delete that file *and* /var/lib/ntp/drift/ntp.drift, then: grep USER_HZ /usr/src/linux/include/asm/param.h adjtimex -p In the latter, the value of "tick" should be 1 million/USER_HZ. If not, then run: adjtimex -t <val> -f 0 -o 0 with <val> equal to that value. Now running "adjtimex -p" again should show that "tick" has that value, while the values of "offset" and "frequency" are zero. The system clock will now be running in an uncorrected, "raw", state. Use "hwclock --hctosys" to bring the system clock into agreement with the hardware clock. Now run "adjtimex -c" to determine any discrepancy between the rates of the system and hardware clocks. Comparisons are 10 seconds apart, and after the first two, adjtimex will suggest values of "tick" and "frequency" that will keep the two clocks at the same rate. Also use "adjtimex -h <timeserver>" at least twice, over a period of at least a couple of hours (the longer the better). Do not reboot the system, do not run the ntp daemon, and do not set any time variables in either system or hardware clocks, during this time to ensure accurate calculations. If /var/log/clocks.log exists, delete it before the first run. Starting with the second run, adjtimex will calculate errors in the frequency, in ppm, for each of the system and hardware clocks. It's probably a good idea to let us have a look at the results of all this at this point. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 22:09, Darryl Gregorash wrote:
On 2006-12-17 00:00, ByteEnable wrote:
<snip> Yeah, I deleted the /etc/adjtime file, played with hwclock, played with ntp, etc. Its just no workie. <sigh>.
Byte
Stop ntpd, delete that file *and* /var/lib/ntp/drift/ntp.drift, then:
grep USER_HZ /usr/src/linux/include/asm/param.h adjtimex -p
In the latter, the value of "tick" should be 1 million/USER_HZ. If not, then run: adjtimex -t <val> -f 0 -o 0
with <val> equal to that value. Now running "adjtimex -p" again should show that "tick" has that value, while the values of "offset" and "frequency" are zero.
The system clock will now be running in an uncorrected, "raw", state. Use "hwclock --hctosys" to bring the system clock into agreement with the hardware clock.
Now run "adjtimex -c" to determine any discrepancy between the rates of the system and hardware clocks. Comparisons are 10 seconds apart, and after the first two, adjtimex will suggest values of "tick" and "frequency" that will keep the two clocks at the same rate.
Also use "adjtimex -h <timeserver>" at least twice, over a period of at least a couple of hours (the longer the better). Do not reboot the system, do not run the ntp daemon, and do not set any time variables in either system or hardware clocks, during this time to ensure accurate calculations. If /var/log/clocks.log exists, delete it before the first run. Starting with the second run, adjtimex will calculate errors in the frequency, in ppm, for each of the system and hardware clocks.
It's probably a good idea to let us have a look at the results of all this at this point.
Darryl: If anything cried out for a wiki page on opensuse this is it. Won't you please consider it? -- _____________________________________ John Andersen
On Sun, 2006-12-17 at 01:09 -0600, Darryl Gregorash wrote:
On 2006-12-17 00:00, ByteEnable wrote:
<snip> Yeah, I deleted the /etc/adjtime file, played with hwclock, played with ntp, etc. Its just no workie. <sigh>.
Byte
Stop ntpd, delete that file *and* /var/lib/ntp/drift/ntp.drift, then:
grep USER_HZ /usr/src/linux/include/asm/param.h
param.h:# define USER_HZ 100 /* .. some user interfaces are in "ticks" */
adjtimex -p mode: 0 offset: 0 frequency: 0 maxerror: 16384000 esterror: 16384000 status: 65 time_constant: 6 precision: 1 tolerance: 33554432 tick: 10000 raw time: 1166350769s 215998us = 1166350769.215998 return value = 5
In the latter, the value of "tick" should be 1 million/USER_HZ. If not, then run: adjtimex -t <val> -f 0 -o 0
with <val> equal to that value. Now running "adjtimex -p" again should show that "tick" has that value, while the values of "offset" and "frequency" are zero.
The system clock will now be running in an uncorrected, "raw", state. Use "hwclock --hctosys" to bring the system clock into agreement with the hardware clock.
Now run "adjtimex -c" to determine any discrepancy between the rates of the system and hardware clocks. Comparisons are 10 seconds apart, and after the first two, adjtimex will suggest values of "tick" and "frequency" that will keep the two clocks at the same rate.
--- current --- -- suggested -- cmos time system-cmos error_ppm tick freq tick freq 1166348614 1.782604 1166348622 3.565470 178286.6 10000 0 1166348630 5.348385 178291.5 10000 0 8217 556300 1166348638 7.131298 178291.3 10000 0 8217 570362 1166348646 8.914216 178291.8 10000 0 8217 537550 1166348654 10.697123 178290.7 10000 0 8217 609425 1166348662 12.480035 178291.2 10000 0 8217 576612 1166348670 14.262944 178290.9 10000 0 8217 596925
Also use "adjtimex -h <timeserver>" at least twice, over a period of at
The estimated error in the cmos clock is 0.710913 +- 0.000014 ppm
least a couple of hours (the longer the better). Do not reboot the system, do not run the ntp daemon, and do not set any time variables in either system or hardware clocks, during this time to ensure accurate calculations. If /var/log/clocks.log exists, delete it before the first run. Starting with the second run, adjtimex will calculate errors in the frequency, in ppm, for each of the system and hardware clocks.
It's probably a good idea to let us have a look at the results of all this at this point.
-- The best way to accelerate a computer running Windows is at 9.81 m/s²
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2006-12-17 03:51, ByteEnable wrote:
On Sun, 2006-12-17 at 01:09 -0600, Darryl Gregorash wrote:
adjtimex -p
mode: 0 offset: 0 frequency: 0 maxerror: 16384000 esterror: 16384000 status: 65 time_constant: 6
I don't know if this will change anything else, but all I've read says that this value is too high for a free-running system clock. You may try "adjtimex -T 0" followed by "adjtimex -c" again, to see if it brings the system clock within reason. I don't expect it to, but I really don't know very much about the internal workings of the sys clock. Everything else is quite OK here.
--- current --- -- suggested -- cmos time system-cmos error_ppm tick freq tick freq 1166348614 1.782604 1166348622 3.565470 178286.6 10000 0
Just so there is no confusion here, without the "-i" parameter, these measurements are taken every 10 seconds of system time, and the error is calculated accordingly. Thus: error = (3.56547 - 1.782604)/10 = 0.1782866, or 178268.8 ppm
Also use "adjtimex -h <timeserver>" at least twice, over a period of at
The estimated error in the cmos clock is 0.710913 +- 0.000014 ppm
This is very good, one of the best hardware clocks I've ever seen. However, it does make me wonder if you are running an over-clocked system. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, 2006-12-17 at 00:00 -0600, ByteEnable wrote:
On Sat, 2006-12-16 at 22:13 -0700, Tom Patton wrote:
Then I have to ask again, have you deleted the /etc/adjtime file?
Yeah, I deleted the /etc/adjtime file, played with hwclock, played with ntp, etc. Its just no workie. <sigh>.
Byte
That's a head scratcher, then. Hope I won't see that when my 10.2 box arrives next week. The last "Time" problem I had was back in 8.2 with dual-boot win98. I'd use win maybe once a month, and then have to chase the adjtime correction for the next 4 days to get it settled down. That's about the time I wiped windows completely, and has been fine since. I'm off about 30 seconds, but haven't set the time in 6 months, and that's after upgrading to 10.1 then removing it. -- Tom in NM SuSE 9.3/Evolution 12:40am up 6 days 11:22, 2 users, load average: 0.15, 0.11, 0.10 ==== -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
ByteEnable wrote:
On Sat, 2006-12-16 at 22:13 -0700, Tom Patton wrote:
On Sat, 2006-12-16 at 21:09 -0600, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Byte
Then I have to ask again, have you deleted the /etc/adjtime file? My understanding is that sets the correction factor for the kernel "system clock". Get it to run correctly before you start messing with NTPD, IMO.
Tom
Yeah, I deleted the /etc/adjtime file, played with hwclock, played with ntp, etc. Its just no workie. <sigh>.
Have you yet tried the boot parameter "clock=pit" as I suggested initially? -- Anders Norrbring Norrbring Consulting
On 2006-12-17 02:49, Anders Norrbring wrote:
Have you yet tried the boot parameter "clock=pit" as I suggested initially?
What exactly does that do, Anders? -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Darryl Gregorash wrote:
On 2006-12-17 02:49, Anders Norrbring wrote:
Have you yet tried the boot parameter "clock=pit" as I suggested initially?
What exactly does that do, Anders?
This is just manually overriding the automatic time source selection. You're using the PIT (programmable interval timer) over what was being automatically selected by tour kernel. If we had a dmesg output of your boot messages we might be able to tell what was going on with your time source. Mark -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Mark Hounschell wrote:
Darryl Gregorash wrote:
On 2006-12-17 02:49, Anders Norrbring wrote:
Have you yet tried the boot parameter "clock=pit" as I suggested initially?
What exactly does that do, Anders?
This is just manually overriding the automatic time source selection. You're using the PIT (programmable interval timer) over what was being automatically selected by tour kernel. If we had a dmesg output of your boot messages we might be able to tell what was going on with your time source.
Mark
more from kernel-parameters.txt: clock= [BUGS=IA-32, HW] gettimeofday clocksource override. [Deprecated] Forces specified clocksource (if avaliable) to be used when calculating gettimeofday(). If specified clocksource is not avalible, it defaults to PIT. Format: { pit | tsc | cyclone | pmtmr } Mark -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Darryl Gregorash wrote:
On 2006-12-17 02:49, Anders Norrbring wrote:
Have you yet tried the boot parameter "clock=pit" as I suggested initially?
What exactly does that do, Anders?
This is just manually overriding the automatic time source selection. You're using the PIT (programmable interval timer) over what was being automatically selected by tour kernel. If we had a dmesg output of your boot messages we might be able to tell what was going on with your time source. OK, thanks (also to jdd for providing that reference, very informative). I did find a relevant line in my boot.msg file so my previous comment
On 2006-12-17 05:57, Mark Hounschell wrote: that nothing of interest is in the syslogs isn't correct -- but I'm not sure it's all that meaningful either. We'll have to wait until Byte tries Anders's suggestion to see if it solves his problem, but a simple "it works" sure won't be overly satisfying (not to me, at any rate -- but I'm a mathematician, not an engineer :-) ). -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Darryl Gregorash a écrit :
On 2006-12-17 02:49, Anders Norrbring wrote:
Have you yet tried the boot parameter "clock=pit" as I suggested initially?
What exactly does that do, Anders?
http://lkml.org/lkml/2004/3/26/222 may be of interest jdd -- http://www.dodin.net http://dodin.org/mediawiki/index.php/GPS_Lowrance_GO -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2006-12-17 02:49, Anders Norrbring wrote:
Have you yet tried the boot parameter "clock=pit" as I suggested initially? What exactly does that do, Anders? I don't know if it is related, but there is this note in the Release notes, "Booting Systems with HPET Timer Problems On some NVIDIA nForce 5 motherboards (in particular from ASUS) or others
Darryl Gregorash wrote: that do not support a HPET timer, proper booting might require the acpi_use_timer_override kernel option at the boot prompt." I realize this is not the same chipset, but trying the boot parameter shouldn't take too long to see if it affects the problem. -- Joe Morris Registered Linux user 231871 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 December 2006 18:09, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Remind me again what kind of processor you have? Don't know if you mentioned it or I just forgot. -- _____________________________________ John Andersen
On Sat, 2006-12-16 at 20:55 -0900, John Andersen wrote:
On Saturday 16 December 2006 18:09, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Remind me again what kind of processor you have? Don't know if you mentioned it or I just forgot.
Pentium 4 @ 3ghZ, i915 chipset. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sunday 17 December 2006 01:05, ByteEnable wrote:
On Sat, 2006-12-16 at 20:55 -0900, John Andersen wrote:
On Saturday 16 December 2006 18:09, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Remind me again what kind of processor you have? Don't know if you mentioned it or I just forgot.
Pentium 4 @ 3ghZ, i915 chipset.
Pentium can cover a lot of territory. Is it a 32bit machine or 64? Does it appear as a single processor or two? -- _____________________________________ John Andersen
On Sun, 2006-12-17 at 00:24 -0900, John Andersen wrote:
On Sunday 17 December 2006 01:05, ByteEnable wrote:
On Sat, 2006-12-16 at 20:55 -0900, John Andersen wrote:
On Saturday 16 December 2006 18:09, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Remind me again what kind of processor you have? Don't know if you mentioned it or I just forgot.
Pentium 4 @ 3ghZ, i915 chipset.
Pentium can cover a lot of territory. Is it a 32bit machine or 64? Does it appear as a single processor or two?
Its a Pentium 4 @ 3.00ghZ/800mhZ with Hyperthreading (HT) 32-bit running on the i915 32-bit chipset. Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sunday 17 December 2006 00:58, ByteEnable wrote:
On Sun, 2006-12-17 at 00:24 -0900, John Andersen wrote:
On Sunday 17 December 2006 01:05, ByteEnable wrote:
On Sat, 2006-12-16 at 20:55 -0900, John Andersen wrote:
On Saturday 16 December 2006 18:09, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Remind me again what kind of processor you have? Don't know if you mentioned it or I just forgot.
Pentium 4 @ 3ghZ, i915 chipset.
Pentium can cover a lot of territory. Is it a 32bit machine or 64? Does it appear as a single processor or two?
Its a Pentium 4 @ 3.00ghZ/800mhZ with Hyperthreading (HT) 32-bit running on the i915 32-bit chipset.
Byte
If Darryl Gregorash's instructions don't solve it, try going into your bios and turning hyperthreading off for an hour to see if it still drifts. -- _____________________________________ John Andersen
On Sun, 2006-12-17 at 00:46 -0900, John Andersen wrote:
On Sunday 17 December 2006 00:58, ByteEnable wrote:
On Sun, 2006-12-17 at 00:24 -0900, John Andersen wrote:
On Sunday 17 December 2006 01:05, ByteEnable wrote:
On Sat, 2006-12-16 at 20:55 -0900, John Andersen wrote:
On Saturday 16 December 2006 18:09, ByteEnable wrote:
My hardware clock, the actual RTC is keeping rock solid time. My computer has a clock on the front panel, that uses the RTC to display the time. That output is perfect. Its only the time in the kernel/OpenSUSE that is inaccurate and running way too fast.
Remind me again what kind of processor you have? Don't know if you mentioned it or I just forgot.
Pentium 4 @ 3ghZ, i915 chipset.
Pentium can cover a lot of territory. Is it a 32bit machine or 64? Does it appear as a single processor or two?
Its a Pentium 4 @ 3.00ghZ/800mhZ with Hyperthreading (HT) 32-bit running on the i915 32-bit chipset.
Byte
If Darryl Gregorash's instructions don't solve it, try going into your bios and turning hyperthreading off for an hour to see if it still drifts.
What will that accomplish? Do you think this is a SMP issue? scheduler? semaphore? spin-lock? what? Byte -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sunday 17 December 2006 01:06, ByteEnable wrote:
What will that accomplish? Do you think this is a SMP issue? scheduler? semaphore? spin-lock? what?
Byte
Don't know for sure, but there have been more than a few smp problems of late. I have a big issue with x86_64 keyboard character duplication which totally goes away If I set my bios to only show one cpu, OR if I lock the X process to a single cpu with taskset. Its an easy way to rule an smp issue out (or in). It would be worth a try, especially if this never happened before on the same hardware, or if it does not currently happen with a live cd from say, Kubuntu or Knoppix or some other distro. -- _____________________________________ John Andersen
On 2006-12-17 03:57, John Andersen wrote:
On Sunday 17 December 2006 01:06, ByteEnable wrote:
What will that accomplish? Do you think this is a SMP issue? scheduler? semaphore? spin-lock? what?
Byte
Don't know for sure, but there have been more than a few smp problems of late.
I have a big issue with x86_64 keyboard character duplication which totally goes away If I set my bios to only show one cpu, OR if I lock the X process to a single cpu with taskset.
Its an easy way to rule an smp issue out (or in).
It would be worth a try, especially if this never happened before on the same hardware, or if it does not currently happen with a live cd from say, Kubuntu or Knoppix or some other distro.
And on that note, I think it would be instructive to know what configurations do have this clock problem. What would be needed? Certainly CPU and kernel version, hyperthreading on/off, what else? -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2006-12-16 02:40, ByteEnable wrote:
Hi,
I'm using OpenSUSE 10.2 and my clock is running too fast. I turned on NTP but NTP only works when its first run, then defaults back to the local clock, which is the clock that is running too fast.
This is a problem specific to 10.2. I've had OpenSUSE 10.1, Fedora Core 5 and 6 on this same hardware without issues.
I can't remember if you had a chance yet (and don't intend to go looking :-) ) to try Anders Norrbring's suggestion. Stealing an idea from what I just read in another, unrelated, thread ("unable to access website...."), have you considered replacing the kernel from 10.2 with the kernel from 10.1? Whatever you do try, given all that has been said in these two threads, perhaps it is time to file a report on Novell's bugzilla. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (17)
-
Anders Norrbring -
ByteEnable -
Carlos E. R. -
Daniel Bauer -
Darryl Gregorash -
Jan Karjalainen -
jdd -
jfweber＠gilweber.com -
Joe Morris (NTM) -
John Andersen -
Joseph Loo -
Leendert Meyer -
Mark Hounschell -
Patrick Shanahan -
Peter Nikolic -
Sandy Drobic -
Tom Patton