Re: ntpd not keeping time
On Wed, Jan 29, 2003 at 02:50:00PM -0500, [EMAIL PROTECTED] wrote: Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. I've tried purging ntp ntp-simple and reinstalling, but that didn't help. There are no errors in any logs in /var/log. Any ideas or suggestions would be greatly appreciated. Thanks. I'm not a hugely experienced user, so I could be way off the mark here, but this reminded me of a problem we had at home a short while ago with a new machine: Basically ntp was setting the time correctly, but the drift was far too large for it to keep up with (over 1 second per minute in our case). In order to solve this, we had to use the adjtimex utility to configure the kernel to expect the actual rate that our system clock was advancing, rather than the default. ( Apparently this is because the master crystal is often uncalibrated, so the clock was advancing at the wrong rate.) After doing this ntp was able to catch up (although adjtimex had to be set to run on every boot, as the change is not permanent). Obviously if it's not a new box I'm seriously far off the mark ;) Regards, Steph -- Stephanie Boyd http://www.ixtab.org.uk/slb The best Windows accelerator is that which works at 9.81m/s2 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
Stephanie Boyd wrote: On Wed, Jan 29, 2003 at 02:50:00PM -0500, [EMAIL PROTECTED] wrote: Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. I've tried purging ntp ntp-simple and reinstalling, but that didn't help. There are no errors in any logs in /var/log. Any ideas or suggestions would be greatly appreciated. Thanks. I'm not a hugely experienced user, so I could be way off the mark here, but this reminded me of a problem we had at home a short while ago with a new machine: Basically ntp was setting the time correctly, but the drift was far too large for it to keep up with (over 1 second per minute in our case). In order to solve this, we had to use the adjtimex utility to configure the kernel to expect the actual rate that our system clock was advancing, rather than the default. ( Apparently this is because the master crystal is often uncalibrated, so the clock was advancing at the wrong rate.) After doing this ntp was able to catch up (although adjtimex had to be set to run on every boot, as the change is not permanent). Unless the time needs to be very exact you could just run ntpdate from cron at suitable interval. (suitable should be relative to the drift speed of your clock). You could also try to run ntpdate just before you start ntpd. Regards, Emil -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
On Fri, Jan 31, 2003 at 04:02:37PM +, Stephanie Boyd wrote: ( Apparently this is because the master crystal is often uncalibrated, so the clock was advancing at the wrong rate.) The hardware RTC relies on a 32.768kHz crystal. The vast majority of 32.768kHz crystals are designed for digital watches, and rely on the wearer's body heat to keep them at a reasonably constant temperature. When they are used in a PC, the temperature is all over the place, therefore so is the crystal frequency, and thus the clock. The other crystals used in a PC don't have such a problem because (a) they're only used when the PC is on and (b) they operate at a much higher frequency, so the chunk of quartz is cut to a totally different shape which is less temperature sensitive. The best Windows accelerator is that which works at 9.81m/s2 Like it... reminds me of *Any* car can be made to do 0-60mph in under 3 seconds - allow me to demonstrate with yours. Pigeon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
Pigeon wrote: [...] Like it... reminds me of *Any* car can be made to do 0-60mph in under 3 seconds - allow me to demonstrate with yours. Achieving ~32 ft/sec^2 acceleration is not that difficult. Finding an unobstructed 121 ft dragstrip with a non-blockaded starting line might be more problematic. -- gt [EMAIL PROTECTED] If you find you've dug yourself into a hole, stop digging. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
# ntptrace clock.psu.edu otc2.psu.edu: stratum 2, offset -529.277338, synch distance 0.06970 ntptrace (and ntpdate) work fine, but ntpd still doesn't sync time. I previously had a very restrictive firewall ruleset on the box, but I've relaxed it quite a bit and still nothing. :( I'm actually considering setting up a cron job to run ntpdate every 15 minutes (on an internal ntp server, of course) until I can replace the motherboard or fix this. Is anyone aware of any weird bugs or clock chipset incompatibilities with ntpd? On Wed, 29 Jan 2003, nate wrote: Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. try 'ntptrace host' sample: defiant:/home/aphro# ntptrace redhat.aphroland.org redhat.aphroland.org: stratum 2, offset 0.001617, synch distance 0.07616 time.nist.gov: stratum 1, offset 0.001895, synch distance 0.0, refid 'ACTS' nate (redhat.aphroland.org is not accessable from the internet so don't bother trying to query it :) ) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
This might be useful to someone out there.. Run as root on the problem system: # ntptime ntp_gettime() returns code 5 (ERROR) time c1e3d62c.cf888000 Thu, Jan 30 2003 11:49:48.810, (.810677), maximum error 16384000 us, estimated error 16 us ntp_adjtime() returns code 5 (ERROR) modes 0x0 (), offset 0.000 us, frequency 0.000 ppm, interval 4 s, maximum error 16384000 us, estimated error 16 us, status 0x41 (PLL,UNSYNC), time constant 0, precision 1.000 us, tolerance 512 ppm, pps frequency 0.000 ppm, stability 512.000 ppm, jitter 200.000 us, intervals 0, jitter exceeded 0, stability exceeded 0, errors 0. Run as root on working systems: # ntptime ntp_gettime() returns code 0 (OK) time c1e3d490.690d1000 Thu, Jan 30 2003 11:42:56.410, (.410356), maximum error 460134 us, estimated error 741 us ntp_adjtime() returns code 0 (OK) modes 0x0 (), offset -26.000 us, frequency 94.460 ppm, interval 4 s, maximum error 460134 us, estimated error 741 us, status 0x1 (PLL), time constant 6, precision 1.000 us, tolerance 512 ppm, pps frequency 0.000 ppm, stability 512.000 ppm, jitter 200.000 us, intervals 0, jitter exceeded 0, stability exceeded 0, errors 0. # ntptime ntp_gettime() returns code 0 (OK) time c1e3d49d.dc3b6000 Thu, Jan 30 2003 11:43:09.860, (.860281), maximum error 356570 us, estimated error 34979 us ntp_adjtime() returns code 0 (OK) modes 0x0 (), offset 7319.000 us, frequency -69.290 ppm, interval 4 s, maximum error 356570 us, estimated error 34979 us, status 0x1 (PLL), time constant 6, precision 1.000 us, tolerance 512 ppm, pps frequency 0.000 ppm, stability 512.000 ppm, jitter 200.000 us, intervals 0, jitter exceeded 0, stability exceeded 0, errors 0. The builds on all systems are nearly identical - configs, packages, etc. So now I'm suspecting some sort of incompatibility between the cheap motherboard I chose and either libc, the kernel, or ntpd. The board, btw, is a Gigabyte GA-7VKML. Or maybe I'm just going wild with speculation again. Help please? :) On Wed, 29 Jan 2003, nate wrote: Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. try 'ntptrace host' sample: defiant:/home/aphro# ntptrace redhat.aphroland.org redhat.aphroland.org: stratum 2, offset 0.001617, synch distance 0.07616 time.nist.gov: stratum 1, offset 0.001895, synch distance 0.0, refid 'ACTS' nate (redhat.aphroland.org is not accessable from the internet so don't bother trying to query it :) ) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: ntpd not keeping time
Ntpdate and ntptrace are over TCP connections. Normal ntpd traffic is over UDP connections, I believe. Check your firewall. Run ntpdc and enter dmpeer dmpeer remote local st poll reach delay offsetdisp === .adsl-65-70-149- 65.70.149.2113 1024 377 0.00063 0.002982 0.01868 *adsl-65-70-149- 65.70.149.2112 1024 377 0.00192 0.000956 0.01485 should show you the some server connect info reach is octal 8 bits which represent connection over the last 8 attempts, i.e., 377 means connected all eight times, 003 means just connected last two times, etc. send the dmpeer output, and I'll try to give you a hand... -Jeff -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Cc: [EMAIL PROTECTED] Subject: Re: ntpd not keeping time # ntptrace clock.psu.edu otc2.psu.edu: stratum 2, offset -529.277338, synch distance 0.06970 ntptrace (and ntpdate) work fine, but ntpd still doesn't sync time. I previously had a very restrictive firewall ruleset on the box, but I've relaxed it quite a bit and still nothing. :( -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
ntpd not keeping time
Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. I've tried purging ntp ntp-simple and reinstalling, but that didn't help. There are no errors in any logs in /var/log. Any ideas or suggestions would be greatly appreciated. Thanks. # uname -a Linux haze 2.4.18 #1 SMP Fri Dec 27 00:01:51 EST 2002 i686 AMD Athlon(tm) Processor AuthenticAMD GNU/Linux # dpkg -l|grep ntp ii ntp4.1.0-8Daemon and utilities for full NTP v4 timekee ii ntp-simple 4.1.0-8NTP v4 daemon for simple systems. ii ntpdate4.1.0-8The ntpdate client for setting system time f # cat /etc/ntp.conf # /etc/ntp.conf, configuration for ntpd # ntpd will use syslog() if logfile is not defined #logfile /var/log/ntpd driftfile /var/lib/ntp/ntp.drift statsdir /var/log/ntpstats/ statistics loopstats peerstats clockstats filegen loopstats file loopstats type day enable filegen peerstats file peerstats type day enable filegen clockstats file clockstats type day enable ### lines starting 'server' are auto generated, ### use dpkg-reconfigure to modify those lines. server timex.cs.columbia.edu server clock.psu.edu server bonehed.lcs.mit.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. try 'ntptrace host' sample: defiant:/home/aphro# ntptrace redhat.aphroland.org redhat.aphroland.org: stratum 2, offset 0.001617, synch distance 0.07616 time.nist.gov: stratum 1, offset 0.001895, synch distance 0.0, refid 'ACTS' nate (redhat.aphroland.org is not accessable from the internet so don't bother trying to query it :) ) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
hi ya turn on logfile in your ntp.conf -- run some ntptrace and ntpdate commands and post its output ntpdate -udb time.apple.com ( use your servers from your ntp.conf file ) ntptrace -dv time.apple.com more ntp jibberish ( urls to other docs too ) http://www.Linux-Consulting.com/NTP/NTP.Commands.txt c ya alvin On Wed, 29 Jan 2003 [EMAIL PROTECTED] wrote: Hello all. I've got this weird problem with ntpd. Up until now I've never had ntpd fail me. But on this one box I've got the clock speeding up by 30 minutes each day. I've got ntpd installed and configured properly with working ntp servers. In fact, ntpdate sets the clock properly, but ntpd is failing to keep the clock in sync with the ntp servers. I've tried purging ntp ntp-simple and reinstalling, but that didn't help. There are no errors in any logs in /var/log. Any ideas or suggestions would be greatly appreciated. Thanks. # uname -a Linux haze 2.4.18 #1 SMP Fri Dec 27 00:01:51 EST 2002 i686 AMD Athlon(tm) Processor AuthenticAMD GNU/Linux # dpkg -l|grep ntp ii ntp4.1.0-8Daemon and utilities for full NTP v4 timekee ii ntp-simple 4.1.0-8NTP v4 daemon for simple systems. ii ntpdate4.1.0-8The ntpdate client for setting system time f # cat /etc/ntp.conf # /etc/ntp.conf, configuration for ntpd # ntpd will use syslog() if logfile is not defined #logfile /var/log/ntpd driftfile /var/lib/ntp/ntp.drift statsdir /var/log/ntpstats/ statistics loopstats peerstats clockstats filegen loopstats file loopstats type day enable filegen peerstats file peerstats type day enable filegen clockstats file clockstats type day enable ### lines starting 'server' are auto generated, ### use dpkg-reconfigure to modify those lines. server timex.cs.columbia.edu server clock.psu.edu server bonehed.lcs.mit.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd not keeping time
Enabling the logfile in ntp.conf doesn't show anything special, just the same start/stop stuff from daemon.log: 29 Jan 16:38:15 ntpd[23209]: frequency initialized 0.000 from /var/lib/ntp/ntp.drift 29 Jan 16:38:15 ntpd[23210]: signal_no_reset: signal 17 had flags 400 29 Jan 16:38:28 ntpd[23209]: ntpd exiting on signal 15 Here are the outputs of those commands: # ntpdate -udb clock.psu.edu 29 Jan 16:38:39 ntpdate[23213]: ntpdate 4.1.0 Mon Mar 25 23:39:50 UTC 2002 (2) transmit(128.118.25.3) transmit(128.118.25.3) transmit(128.118.25.3) transmit(128.118.25.3) transmit(128.118.25.3) server 128.118.25.3, port 123 stratum 0, precision 0, leap 00, trust 000 refid [0.0.0.0], delay 0.0, dispersion 64.0 transmitted 4, in filter 4 reference time:. Thu, Feb 7 2036 1:28:16.000 originate timestamp: . Thu, Feb 7 2036 1:28:16.000 transmit timestamp: c1e2c862.691c6d1e Wed, Jan 29 2003 16:38:42.410 filter delay: 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 filter offset: 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 delay 0.0, dispersion 64.0 offset 0.00 29 Jan 16:38:43 ntpdate[23213]: no server suitable for synchronization found # ntptrace -dv clock.psu.edu DoTransmit(128.118.25.3) DoTransmit to 128.118.25.3 timeout DoTransmit(128.118.25.3) DoTransmit to 128.118.25.3 timeout DoTransmit(128.118.25.3) DoTransmit to 128.118.25.3 timeout DoTransmit(128.118.25.3) DoTransmit to 128.118.25.3 timeout DoTransmit(128.118.25.3) DoTransmit to 128.118.25.3 timeout otc2.psu.edu: *Timeout* And here's just plain 'ole ntpdate: # ntpdate clock.psu.edu 29 Jan 16:39:56 ntpdate[23274]: step time server 128.118.25.3 offset -55.557580 sec So it looks like running ntpdate without options works, while running with those options does not. I've double-checked the firewall ruleset and it is exactly the same as other boxes with ntp running just fine. For some reason, though, on the problem box I found this: Jan 29 16:40:30 haze kernel: Packet log: input DENY eth0 PROTO=17 128.118.25.3:123 64.61.10.115:32962 L=76 S=0 So I changed the rules for ntp hosts to allow udp:123 to anywhere and now the above commands return this: # ntpdate -udb clock.psu.edu 29 Jan 17:41:21 ntpdate[24452]: ntpdate 4.1.0 Mon Mar 25 23:39:50 UTC 2002 (2) transmit(128.118.25.3) receive(128.118.25.3) transmit(128.118.25.3) receive(128.118.25.3) transmit(128.118.25.3) receive(128.118.25.3) transmit(128.118.25.3) receive(128.118.25.3) transmit(128.118.25.3) server 128.118.25.3, port 123 stratum 2, precision -16, leap 00, trust 000 refid [192.5.41.209], delay 0.04216, dispersion 0.00023 transmitted 4, in filter 4 reference time:c1e2d574.ef5cf249 Wed, Jan 29 2003 17:34:28.935 originate timestamp: c1e2d6f3.a58ab92c Wed, Jan 29 2003 17:40:51.646 transmit timestamp: c1e2d711.43e64f54 Wed, Jan 29 2003 17:41:21.265 filter delay: 0.04216 0.04269 0.04330 0.04219 0.0 0.0 0.0 0.0 filter offset: -29.6269 -29.6268 -29.6273 -29.6272 0.00 0.00 0.00 0.00 delay 0.04216, dispersion 0.00023 offset -29.626950 29 Jan 17:41:21 ntpdate[24452]: step time server 128.118.25.3 offset -29.626950 sec # ntptrace -dv clock.psu.edu DoTransmit(128.118.25.3) DoTransmit to 128.118.25.3 ReceiveBuf(128.118.25.3, 128.118.25.3) server 128.118.25.3, port 123 stratum 2, precision -16, leap 00 refid ntp2.usno.navy.mil delay 0.01788, dispersion 0.0 offset -29.872159 rootdelay 0.03152, rootdispersion 0.04436, synch dist 0.06012 reference time: c1e2d574.ef5cf249 Wed, Jan 29 2003 17:34:28.935 originate timestamp: c1e2d711.b211ec91 Wed, Jan 29 2003 17:41:21.695 transmit timestamp: c1e2d72f.8efc9bc7 Wed, Jan 29 2003 17:41:51.558 DoTransmit(192.5.41.209) DoTransmit to 192.5.41.209 timeout DoTransmit(192.5.41.209) DoTransmit to 192.5.41.209 timeout DoTransmit(192.5.41.209) DoTransmit to 192.5.41.209 timeout DoTransmit(192.5.41.209) DoTransmit to 192.5.41.209 timeout DoTransmit(192.5.41.209) DoTransmit to 192.5.41.209 timeout ntp2.usno.navy.mil: *Timeout* So now nothing is being blocked by the firewall (except for ntp2.usno.navy.mil) and ntp still isn't sync'ing time on the box. I even tried disabling the firewall for a few minutes, still no time sync. In a few hours the clock has managed to creep ahead by 30 seconds. I'm wondering if something's up with the clock on the motherboard, some freaky incompatibility with the way ntpd updates time or something. It's a Gigabyte GA-7VKML, in case anyone's curious. Just a possibility. Or maybe when ntp calculates the drift or how long it takes to get a response from the time server, something is causing it to miscalculate each time which could be causing it to slowly increase it's adjustments. Or maybe I'm just going wild with speculation. :) On Wed, 29 Jan 2003, Alvin Oga wrote: hi ya turn on logfile in your ntp.conf -- run some
Re: ntpd not keeping time -fw
hi ya debian since ntpdate does sync w/ clock.psu.edu and xntpd does not, you will need to fix your firewall rules c ya alvin On Wed, 29 Jan 2003 [EMAIL PROTECTED] wrote: # ntpdate -udb clock.psu.edu ... 29 Jan 16:38:43 ntpdate[23213]: no server suitable for synchronization found # ntptrace -dv clock.psu.edu ... otc2.psu.edu: *Timeout* ... And here's just plain 'ole ntpdate: # ntpdate clock.psu.edu 29 Jan 16:39:56 ntpdate[23274]: step time server 128.118.25.3 offset -55.557580 sec .. So it looks like running ntpdate without options works, while running with those options does not. I've double-checked the firewall ruleset and it is exactly the same as other boxes with ntp running just fine. deesnt mean that its right for that box.. kernel issues, ipchains/iptable issues, xntp versions, etc For some reason, though, on the problem box I found this: Jan 29 16:40:30 haze kernel: Packet log: input DENY eth0 PROTO=17 128.118.25.3:123 64.61.10.115:32962 L=76 S=0 So I changed the rules for ntp hosts to allow udp:123 to anywhere and now the above commands return this: .. # ntpdate -udb clock.psu.edu .. 29 Jan 17:41:21 ntpdate[24452]: step time server 128.118.25.3 offset -29.626950 sec # ntptrace -dv clock.psu.edu ... refid ntp2.usno.navy.mil delay 0.01788, dispersion 0.0 offset -29.872159 ... So now nothing is being blocked by the firewall (except for ntp2.usno.navy.mil) and ntp still isn't sync'ing time on the box. I even you have to wait ... say several hours to see if you can get regular log messages in /var/log/xntpd - mine defaults to once/hr for xntp sync'ing c ya alvin .. more ntp jibberish ( urls to other docs too ) http://www.Linux-Consulting.com/NTP/NTP.Commands.txt -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
