Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
64 bytes from 87.186.242.38: icmp_seq=1 ttl=254 time=48.3 ms 64 bytes from 87.186.242.38: icmp_seq=1 ttl=245 time=34.7 ms Reply from 87.186.242.38: bytes=32 time=43ms TTL=239 I like the different TTL values, showing how many hops there are between .38 and the folks pinging it. Based on the first one, it seems reasonable to assume that .38 sends the replies with a TTL of 255. The variation were from my daughter using the connection for watching youtube in parallel. ( NAT hidden home network. ) http://www.bufferbloat.net/ and what they are doing might be of interest to you. rick jones -- portable adj, code that compiles under more than one compiler these opinions are mine, all mine; HP might not want them anyway... :) feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH... ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Terje Mathisen terje.mathisen at tmsw.no wrote: unruh wrote: And it sounds like you have assymetric delays. Note that most ISPs deliver very different rates for up vs down, and that may well come with assymetric delays. (eg 600Kb/s, vs 30Mb/s for my cable access) Ouch! That is 1:50 which means that you must be very lucky indeed to ever get those 30 Mbit down: The most efficient protocol for bulk transfers is still FTP, if you get zero packet drops FTP can run with one ACK packet (about 64 bytes) for every pair of received data packets (2x1514 max). The ratio between those two numbers is just over 1:50, so in order to ever get nearly the full 30 Mbit down, you must have perfect saturation of the uplink just to send the ACK packets, and any background traffic, even something as simple as NTP of email polling will interfere. You should never accept much more than 10:1 speed difference between down and up. I will agree with that, but will also point-out there are some stacks with explicit ACK avoidance heuristics (Solaris and HP-UX), and that Large Receive Offload/GRO in Linux will cause what I believe are called stretch ACKs that take one beyond ack-every-other to ack-every-several-or-many. With the prevalence of timestamps being enabled in TCP, I think a TCP over Ethernet will be larger than the minimum frame size - 32 bytes of TCP header (12 bytes of timestamp, not sure about padding/alignment), 20 bytes of IPv4 header, and 14 bytes of ethernet header, for a total of 66 bytes. Make that IPv6 and add another 20 bytes to arrive at 86, IIRC. rick jones -- The glass is neither half-empty nor half-full. The glass has a leak. The real question is Can it be patched? these opinions are mine, all mine; HP might not want them anyway... :) feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH... ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) wrote: On 3/17/2012 6:00 PM, Terje Mathisen wrote: I prefer fiber: 50/50 Mbit, very consistent ping times of a couple of ms to most no.pool.ntp.org servers. :-) Cost is about $75/month. Terje You're a lucky dude if that's even an option. For most purposes other than timing, Comcast cable does pretty good for me. I've even had fiber in my mountain cabin since day one, 5-7 years ago. (Check http://visitrauland.com , we're in the middle of xc trails!) Terje -- - Terje.Mathisen at tmsw.no almost all programming can be viewed as an exercise in caching ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
David J Taylor wrote: John Hasler jhas...@newsguy.com wrote in message news:87sjh7mlqs@thumper.dhh.gt.org... David J Taylor writes: But in the UK from Virgin Media I have 30 Mb/s down, 1 Mb/s up. I have been promised an upload speed increase about 18 months ago to 2 Mb/s up, which is more sensible... Such a very high cable download speed is a peak burst speeds on a shared medium. Your sustained performance is not likely to be more than a fraction of it. Speed tests from a number of sites show 30 Mb/s, and that's over several seconds. I downloaded the Windows-8 64-bit ISO recently, and the 3,583,707,136 bytes took 48 minutes, 24 seconds, which I make about 7.4 Mb/s, and the 32-bit ISO was 2,711,396,352 bytes in 27 minutes, so about 13.4 Mb/s. That was without any download accelerator (no multiple connections). Sustained transfer speeds are a moot point. You want to know about chanel delays. In absolute, asymmetricness and variation terms. Regular DSL here has quite large and spread line delays though speed is much higher delay is similar or slightly larger than forex ISDN. PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. ( my first pingable outside node ) 64 bytes from 87.186.242.38: icmp_seq=1 ttl=254 time=48.3 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=254 time=34.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=254 time=77.4 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=254 time=70.8 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=254 time=108 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=254 time=89.0 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=254 time=109 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=254 time=64.1 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=254 time=76.1 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=254 time=145 ms 64 bytes from 87.186.242.38: icmp_seq=11 ttl=254 time=199 ms 64 bytes from 87.186.242.38: icmp_seq=12 ttl=254 time=104 ms 64 bytes from 87.186.242.38: icmp_seq=13 ttl=254 time=247 ms 64 bytes from 87.186.242.38: icmp_seq=14 ttl=254 time=104 ms uwe ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Uwe Klein u...@klein-habertwedt.de wrote: Regular DSL here has quite large and spread line delays though speed is much higher delay is similar or slightly larger than forex ISDN. PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. ( my first pingable outside node ) 64 bytes from 87.186.242.38: icmp_seq=1 ttl=254 time=48.3 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=254 time=34.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=254 time=77.4 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=254 time=70.8 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=254 time=108 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=254 time=89.0 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=254 time=109 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=254 time=64.1 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=254 time=76.1 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=254 time=145 ms 64 bytes from 87.186.242.38: icmp_seq=11 ttl=254 time=199 ms 64 bytes from 87.186.242.38: icmp_seq=12 ttl=254 time=104 ms 64 bytes from 87.186.242.38: icmp_seq=13 ttl=254 time=247 ms 64 bytes from 87.186.242.38: icmp_seq=14 ttl=254 time=104 ms Funny network... I can ping the same address over my own DSL and get lower and more stable ping than you do: ping 87.186.242.38 PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. 64 bytes from 87.186.242.38: icmp_seq=1 ttl=245 time=34.7 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=245 time=36.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=245 time=35.3 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=245 time=33.6 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=245 time=34.8 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=245 time=35.8 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=245 time=34.2 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=245 time=35.3 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=245 time=33.0 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=245 time=34.2 ms ^C --- 87.186.242.38 ping statistics --- 10 packets transmitted, 10 received, 0% packet loss, time 9043ms rtt min/avg/max/mdev = 33.069/34.776/36.339/0.943 ms Pinging something local in my provider yields stable pingtimes within 13-14 ms... Maybe your provider still uses old ATM technology between the subscribers and the DSL router, and the network is heavily overbooked. This is, however, not a generic property of DSL. DSL can have stable roundtrip times. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
I prefer fiber: 50/50 Mbit, very consistent ping times of a couple of ms to most no.pool.ntp.org servers. :-) Cost is about $75/month. Terje Maybe I should move to Norway! You are lucky! I pay US $48 (equivalent) for my 30/1 service! Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f65006d.4090...@c3energy.com... [] Come to think of it, my comment about the polling interval not increasing may only apply to a local refclock, not a local server. You may be right - all the servers on this test system are Internet servers. Can you elaborate more about what NTPD_USE_INTERP_DANGEROUS does? Sincerely, Ron Overnight results are here: http://www.satsignal.eu/ntp/Win-8+Internet.html As you can see, jitter is under a millisecond, with the local (noselect) server showing about half a millisecond of jitter, and the Internet servers between about 3 and just over 6 milliseconds. NTPD_USE_INTERP_DANGEROUS if set, forces interpolation on Windows. With Windows XP or earlier, Interpolation works well, but not on Vista or later, so under certain conditions, NTP disables interpolation as mentioned here: http://lists.ntp.org/pipermail/questions/2011-November/030904.html You can enable interpolation, and you may get better results. If not, jitters will have a floor of about 0.977 milliseconds. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Rob nom...@example.com wrote in message news:slrnjmbdn3.9ce.nom...@xs8.xs4all.nl... Uwe Klein u...@klein-habertwedt.de wrote: Regular DSL here has quite large and spread line delays though speed is much higher delay is similar or slightly larger than forex ISDN. PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. ( my first pingable outside node ) 64 bytes from 87.186.242.38: icmp_seq=1 ttl=254 time=48.3 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=254 time=34.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=254 time=77.4 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=254 time=70.8 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=254 time=108 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=254 time=89.0 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=254 time=109 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=254 time=64.1 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=254 time=76.1 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=254 time=145 ms 64 bytes from 87.186.242.38: icmp_seq=11 ttl=254 time=199 ms 64 bytes from 87.186.242.38: icmp_seq=12 ttl=254 time=104 ms 64 bytes from 87.186.242.38: icmp_seq=13 ttl=254 time=247 ms 64 bytes from 87.186.242.38: icmp_seq=14 ttl=254 time=104 ms Funny network... I can ping the same address over my own DSL and get lower and more stable ping than you do: ping 87.186.242.38 PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. 64 bytes from 87.186.242.38: icmp_seq=1 ttl=245 time=34.7 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=245 time=36.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=245 time=35.3 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=245 time=33.6 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=245 time=34.8 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=245 time=35.8 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=245 time=34.2 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=245 time=35.3 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=245 time=33.0 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=245 time=34.2 ms ^C --- 87.186.242.38 ping statistics --- 10 packets transmitted, 10 received, 0% packet loss, time 9043ms rtt min/avg/max/mdev = 33.069/34.776/36.339/0.943 ms Pinging something local in my provider yields stable pingtimes within 13-14 ms... Maybe your provider still uses old ATM technology between the subscribers and the DSL router, and the network is heavily overbooked. This is, however, not a generic property of DSL. DSL can have stable roundtrip times. .. and from Edinburgh: ping 87.186.242.38 Pinging 87.186.242.38 with 32 bytes of data: Reply from 87.186.242.38: bytes=32 time=43ms TTL=239 Reply from 87.186.242.38: bytes=32 time=44ms TTL=239 Reply from 87.186.242.38: bytes=32 time=46ms TTL=239 Reply from 87.186.242.38: bytes=32 time=50ms TTL=239 Ping statistics for 87.186.242.38: Packets: Sent = 4, Received = 4, Lost = 0 (0% loss), Approximate round trip times in milli-seconds: Minimum = 43ms, Maximum = 50ms, Average = 45ms That's at 11:07 on a Sunday morning. David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
David J Taylor wrote: Rob nom...@example.com wrote in message news:slrnjmbdn3.9ce.nom...@xs8.xs4all.nl... Uwe Klein u...@klein-habertwedt.de wrote: Regular DSL here has quite large and spread line delays though speed is much higher delay is similar or slightly larger than forex ISDN. PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. ( my first pingable outside node ) 64 bytes from 87.186.242.38: icmp_seq=1 ttl=254 time=48.3 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=254 time=34.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=254 time=77.4 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=254 time=70.8 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=254 time=108 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=254 time=89.0 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=254 time=109 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=254 time=64.1 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=254 time=76.1 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=254 time=145 ms 64 bytes from 87.186.242.38: icmp_seq=11 ttl=254 time=199 ms 64 bytes from 87.186.242.38: icmp_seq=12 ttl=254 time=104 ms 64 bytes from 87.186.242.38: icmp_seq=13 ttl=254 time=247 ms 64 bytes from 87.186.242.38: icmp_seq=14 ttl=254 time=104 ms Funny network... I can ping the same address over my own DSL and get lower and more stable ping than you do: ping 87.186.242.38 PING 87.186.242.38 (87.186.242.38) 56(84) bytes of data. 64 bytes from 87.186.242.38: icmp_seq=1 ttl=245 time=34.7 ms 64 bytes from 87.186.242.38: icmp_seq=2 ttl=245 time=36.3 ms 64 bytes from 87.186.242.38: icmp_seq=3 ttl=245 time=35.3 ms 64 bytes from 87.186.242.38: icmp_seq=4 ttl=245 time=33.6 ms 64 bytes from 87.186.242.38: icmp_seq=5 ttl=245 time=34.8 ms 64 bytes from 87.186.242.38: icmp_seq=6 ttl=245 time=35.8 ms 64 bytes from 87.186.242.38: icmp_seq=7 ttl=245 time=34.2 ms 64 bytes from 87.186.242.38: icmp_seq=8 ttl=245 time=35.3 ms 64 bytes from 87.186.242.38: icmp_seq=9 ttl=245 time=33.0 ms 64 bytes from 87.186.242.38: icmp_seq=10 ttl=245 time=34.2 ms ^C --- 87.186.242.38 ping statistics --- 10 packets transmitted, 10 received, 0% packet loss, time 9043ms rtt min/avg/max/mdev = 33.069/34.776/36.339/0.943 ms Pinging something local in my provider yields stable pingtimes within 13-14 ms... Maybe your provider still uses old ATM technology between the subscribers and the DSL router, and the network is heavily overbooked. This is, however, not a generic property of DSL. DSL can have stable roundtrip times. .. and from Edinburgh: ping 87.186.242.38 Pinging 87.186.242.38 with 32 bytes of data: Reply from 87.186.242.38: bytes=32 time=43ms TTL=239 Reply from 87.186.242.38: bytes=32 time=44ms TTL=239 Reply from 87.186.242.38: bytes=32 time=46ms TTL=239 Reply from 87.186.242.38: bytes=32 time=50ms TTL=239 Ping statistics for 87.186.242.38: Packets: Sent = 4, Received = 4, Lost = 0 (0% loss), Approximate round trip times in milli-seconds: Minimum = 43ms, Maximum = 50ms, Average = 45ms That's at 11:07 on a Sunday morning. David Deutsche Telekom. Best speed optimised over 4km of copper wire. The variation were from my daughter using the connection for watching youtube in parallel. ( NAT hidden home network. ) uwe ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/18/2012 04:03, David J Taylor wrote: I prefer fiber: 50/50 Mbit, very consistent ping times of a couple of ms to most no.pool.ntp.org servers. :-) Cost is about $75/month. Terje Maybe I should move to Norway! You are lucky! I pay US $48 (equivalent) for my 30/1 service! Don't complain too much. I pay $78 for 25/2 though I do get a /29 netblock (netmask 255.255.255.248) out of it. Eight public IPs, five usable after broadcast, gateway, and route are taken out. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
unruh wrote: Would it be worth it to recruit an electrical or systems engineer who claimed to know something about filtering data to take a serious look at NTPD's data filtering approach? There has to be some reason that there is a David Mills claims to know about filtering data. Not that I always agree with him, but he is not stupid. That must be the understatement of the day... And it sounds like you have assymetric delays. Note that most ISPs deliver very different rates for up vs down, and that may well come with assymetric delays. (eg 600Kb/s, vs 30Mb/s for my cable access) Ouch! That is 1:50 which means that you must be very lucky indeed to ever get those 30 Mbit down: The most efficient protocol for bulk transfers is still FTP, if you get zero packet drops FTP can run with one ACK packet (about 64 bytes) for every pair of received data packets (2x1514 max). The ratio between those two numbers is just over 1:50, so in order to ever get nearly the full 30 Mbit down, you must have perfect saturation of the uplink just to send the ACK packets, and any background traffic, even something as simple as NTP of email polling will interfere. You should never accept much more than 10:1 speed difference between down and up. Terje -- - Terje.Mathisen at tmsw.no almost all programming can be viewed as an exercise in caching ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Terje Mathisen terje.mathisen at tmsw.no wrote in message news:is9e39-6ve2@ntp6.tmsw.no... [] You should never accept much more than 10:1 speed difference between down and up. Terje Should - I agree. But in the UK from Virgin Media I have 30 Mb/s down, 1 Mb/s up. I have been promised an upload speed increase about 18 months ago to 2 Mb/s up, which is more sensible, but that increase hasn't yet been delivered. I am also promised an increase within the next year which doubles my existing speeds, giving 60 Mb/s down and 4 Mb/s up. Not ideal, but I prefer cable to ADSL. No wonder I have four GPS devices connected to 5 PCs! G Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
David J Taylor writes: But in the UK from Virgin Media I have 30 Mb/s down, 1 Mb/s up. I have been promised an upload speed increase about 18 months ago to 2 Mb/s up, which is more sensible... Such a very high cable download speed is a peak burst speeds on a shared medium. Your sustained performance is not likely to be more than a fraction of it. Not ideal, but I prefer cable to ADSL. You'd get less jitter with DSL. -- John Hasler jhas...@newsguy.com Dancing Horse Hill Elmwood, WI USA ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
John Hasler jhas...@newsguy.com wrote in message news:87sjh7mlqs@thumper.dhh.gt.org... David J Taylor writes: But in the UK from Virgin Media I have 30 Mb/s down, 1 Mb/s up. I have been promised an upload speed increase about 18 months ago to 2 Mb/s up, which is more sensible... Such a very high cable download speed is a peak burst speeds on a shared medium. Your sustained performance is not likely to be more than a fraction of it. Speed tests from a number of sites show 30 Mb/s, and that's over several seconds. I downloaded the Windows-8 64-bit ISO recently, and the 3,583,707,136 bytes took 48 minutes, 24 seconds, which I make about 7.4 Mb/s, and the 32-bit ISO was 2,711,396,352 bytes in 27 minutes, so about 13.4 Mb/s. That was without any download accelerator (no multiple connections). Not ideal, but I prefer cable to ADSL. You'd get less jitter with DSL. -- John Hasler DSL or ADSL? But I haven't checked the jitter on the cable modem link for a long time, since I installed the GPS receivers. Thanks, John. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
You'd get less jitter with DSL. -- John Hasler John, You have piqued my interest. I have just set up a Windows-8 PC with an ntp configuration not dissimilar to Ron's, in that it's using purely Internet servers but trying to monitor a local stratum-1 server as well. In essence: __ # Local stratum-1 Free BSD server server 192.168.0.3 iburst minpoll 5 maxpoll 5 prefer noselect # Seven external servers: server x.x.x.uk iburst server y.y.y.uk iburst server 0.uk.pool.ntp.org iburst server 1.uk.pool.ntp.org iburst server 2.uk.pool.ntp.org iburst server 0.nl.pool.ntp.org iburst server 1.nl.pool.ntp.org iburst __ The performance will appear here: http://www.satsignal.eu/mrtg/performance_torvik.php It will be interesting to see whether the Internet servers poll period increased from 64s (I'm hoping that the 32s poll on the local server won't affect the Internet ones), and what level of jitter is achieved. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
I did not express myself well. Permit me to explain. 1. The variable delays are caused by buffering at the Internet routers. See CACM Staff. (2012). BufferBloat: what's wrong with the internet? Commun. ACM, 55(2), 40-47. doi: 10.1145/2076450.2076464. A discussion with Vint Cerf, Van Jacobson, Nick Weaver, and Jim Gettys. Instead of rejecting packets that a router has no bandwidth to transmit, routers now buffer them. As memory has become cheaper, the buffers have in many cases become large. This results in packet delays that are highly variable and can exceed several seconds. The delays are asymmetric; that is, they are always positive. NTPD's filtering algorithm always rejects outliers. So when a packet is received that has been buffered for 500 to 1000 ms, NTPD may associate it with an offset this is much closer to the mean. If this is true, then NTPD has to be biased in its estimation of round-trip delay. If a series of packets are received that have been buffered, then eventually the mean will increase until the offset is no longer an outlier. But if the delay caused by buffering is sporadic, as Unruh suggests, then NTPD will never compute the offset correctly as all the outliers will have been rejected. Furthermore, there are two bits of information associated with the computation of offset: The offset = ((t1-t0) + (t3 - t2))/2 itself and the time it is computed, which must be separated by some power of 2 seconds. As far as I can tell NTPD just throws away the fact that the interval between offset computations must be near some power of 2. Let the offset be y and the computation of it be x. Then is not y = f(x)? That is, is not the offset at least partially a function of the time it was computed? Would regression analysis be a more appropriate filtering algorithm? Moreover, in industry feedback control is often the mechanism used to correct for random noise, which buffering delay may be. 2. It is true that ADSL is asymmetric. For example several tests show that I see [2.57, 2.88] mbps down and [0.560, 0.720] mbps up. But consider that I have to be within a few thousand feet of the telco central office, which is in Center City Philadelphia, and all the NTP servers are in State College, PA (152 mi), Northern (Ramsey) New Jersey (95 mi), or New York City (86 mi). Thus the distance that the packet travels at 0.7 mbps is relatively tiny compared to its total distance. I question whether the ADSL asymmetry accounts for the large negative correlation between delay and offset that NTPD produces. 3. Ron is correct about the BU-353 GPS device: After a cold restart, the BU-353 is right in synch with the average of 8 stratum 2 servers, and then the offset steadily increases until the BU-353 looses synch with the satellites after about a day. Another cold restart starts the cycle again. This may make comparisons between the BU-353 and Internet time invalid. The question is, what would happen if one did a warm restart instead of a cold one? Charles Elliott -Original Message- From: questions-bounces+elliott.ch=verizon@lists.ntp.org [mailto:questions-bounces+elliott.ch=verizon@lists.ntp.org] On Behalf Of unruh Sent: Friday, March 16, 2012 7:56 PM To: questions@lists.ntp.org Subject: Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error. On 2012-03-16, Charles Elliott elliott...@verizon.net wrote: On the subject of accuracy, has anyone ever really looked at NTPD's offset filtering mechanism? What it does now is sort the last (about 50) offsets from smallest to largest and then prunes the smallest or largest, depending on which is further away from the average, until there are only N (I forget what N is) offset observations left. That is for refclocks, And it is usually about 16 (poll 4, and once per second). N is about 60% of the total. There may be at least two problems with this filtering mechanism. First, there is no apparent theory behind it; I have never seen such a crude filter The theory is that there are two noise mechanisms, one approximately gaussian with small standard deviation and one much broader but rarer. Ie, occasionally you will get popconr spikes. The median is the optimal estimator if you want to minimize |y-ybar|, just as the mean is the optimal estimator for (y-ybar)^2. |y-ybar| is less sensitive to large deviations. that does not take into account any information inherent in the data. On the other hand, what I don't know about filters would fill all 24 volumes of an encyclopedia. Sure it does. See above. Second, we know that each offset observation should have arrived about one second after the previous one, yet NTPD does not take advantage of that knowledge. There are filters, such as the Kalman filter that uses a Bayesian estimation approach to predict the next observation and adjusts it according to the prediction when it arrives, that do
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/17/2012 11:48 AM, David J Taylor wrote: You'd get less jitter with DSL. -- John Hasler John, You have piqued my interest. I have just set up a Windows-8 PC with an ntp configuration not dissimilar to Ron's, in that it's using purely Internet servers but trying to monitor a local stratum-1 server as well. In essence: __ # Local stratum-1 Free BSD server server 192.168.0.3 iburst minpoll 5 maxpoll 5 prefer noselect # Seven external servers: server x.x.x.uk iburst server y.y.y.uk iburst server 0.uk.pool.ntp.org iburst server 1.uk.pool.ntp.org iburst server 2.uk.pool.ntp.org iburst server 0.nl.pool.ntp.org iburst server 1.nl.pool.ntp.org iburst __ The performance will appear here: http://www.satsignal.eu/mrtg/performance_torvik.php It will be interesting to see whether the Internet servers poll period increased from 64s (I'm hoping that the 32s poll on the local server won't affect the Internet ones), and what level of jitter is achieved. Cheers, David Hi David, I'm not sure what will happen if you simultaneously prefer and noselect the local server. Assuming the local stratum 1 server is the most stable time source, you'll get a much better picture of what the internet servers are doing relative to it if you allow it to be selectable as well as being preferred. When you graph it, if the local server is the active clock, all the lines for the internet servers will be gathered around and relative to the local server. When I tried to do things the other way around, with an internet server preferred, the graph looked awful because there was so much variation. Also, if your local server starts reporting time that looks too far from the internet servers, regardless of who's fault it is, ntp will clock hop over to the internet servers. I don't THINK your internet servers will ever poll above their default minpoll value of 6, or 64 seconds. I realize you don't have a gps attached to this pc, but the iburst lines reminded me of something. I read somewhere that having iburst on internet server lines, if a local gps is attached, could prevent the PC from synchronizing to the gps before it synchronizes to the internet. On my pc with the gps attached, I don't use the iburst command. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f64d793.9010...@c3energy.com... [] Hi David, I'm not sure what will happen if you simultaneously prefer and noselect the local server. Assuming the local stratum 1 server is the most stable time source, you'll get a much better picture of what the internet servers are doing relative to it if you allow it to be selectable as well as being preferred. When you graph it, if the local server is the active clock, all the lines for the internet servers will be gathered around and relative to the local server. When I tried to do things the other way around, with an internet server preferred, the graph looked awful because there was so much variation. Also, if your local server starts reporting time that looks too far from the internet servers, regardless of who's fault it is, ntp will clock hop over to the internet servers. I don't THINK your internet servers will ever poll above their default minpoll value of 6, or 64 seconds. I realize you don't have a gps attached to this pc, but the iburst lines reminded me of something. I read somewhere that having iburst on internet server lines, if a local gps is attached, could prevent the PC from synchronizing to the gps before it synchronizes to the internet. On my pc with the gps attached, I don't use the iburst command. Sincerely, Ron Ron, The local stratum-1 server shows without a tally code against it in the ntpq -p output, so it's being recorded in the peerstats, but not used for syncing. The noselect msut override the prefer. After about three hours running, the Internet servers are all at 512 seconds poll interval. The averaged jitter has been below 1 millisecond for the last couple of hours. The offset is reporting between -0.7 and -1.8 milliseconds, and the frequency is stabilising very nicely (because of the long poll interval). I'll leave this running overnight and tomorrow to see how it handles temperature changes and any Internet access changes, and to get a few more points on the graph. One caveat is that I am using the most recent NTP (ntpd 4.2.7p263 from Dave Hart's download page), and that with Windows-8, it may be using the new precision time system call. From my own tests, this is similar, on earlier versions of NTP, to setting the environment variable NTPD_USE_INTERP_DANGEROUS, thus forcing the NTP time interpolation to be used. The configuration I have is: - cable modem (with built-in router, but working as a bridge by putting my own router in a device in the DMZ). - Samknows network monitor (modified WRT54GL router) - WRT54GL router running DD-WRT firmware - Netgear 8-port consumer 1 Gb/s switch (G5108) - wired connection to ~2 year old laptop PC I only mention this to show that (a) it's not a direct connection and (b) there's no wireless involved. My aim here is simply to see what performance may be had with just Internet servers. The PC is only running NTP and monitoring software - no user programs and no interactive work, so it is a best-case scenario. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/17/2012 3:20 PM, David J Taylor wrote: Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f64d793.9010...@c3energy.com... [] Hi David, I'm not sure what will happen if you simultaneously prefer and noselect the local server. Assuming the local stratum 1 server is the most stable time source, you'll get a much better picture of what the internet servers are doing relative to it if you allow it to be selectable as well as being preferred. When you graph it, if the local server is the active clock, all the lines for the internet servers will be gathered around and relative to the local server. When I tried to do things the other way around, with an internet server preferred, the graph looked awful because there was so much variation. Also, if your local server starts reporting time that looks too far from the internet servers, regardless of who's fault it is, ntp will clock hop over to the internet servers. I don't THINK your internet servers will ever poll above their default minpoll value of 6, or 64 seconds. I realize you don't have a gps attached to this pc, but the iburst lines reminded me of something. I read somewhere that having iburst on internet server lines, if a local gps is attached, could prevent the PC from synchronizing to the gps before it synchronizes to the internet. On my pc with the gps attached, I don't use the iburst command. Sincerely, Ron Ron, The local stratum-1 server shows without a tally code against it in the ntpq -p output, so it's being recorded in the peerstats, but not used for syncing. The noselect msut override the prefer. After about three hours running, the Internet servers are all at 512 seconds poll interval. The averaged jitter has been below 1 millisecond for the last couple of hours. The offset is reporting between -0.7 and -1.8 milliseconds, and the frequency is stabilising very nicely (because of the long poll interval). I'll leave this running overnight and tomorrow to see how it handles temperature changes and any Internet access changes, and to get a few more points on the graph. Come to think of it, my comment about the polling interval not increasing may only apply to a local refclock, not a local server. One caveat is that I am using the most recent NTP (ntpd 4.2.7p263 from Dave Hart's download page), and that with Windows-8, it may be using the new precision time system call. From my own tests, this is similar, on earlier versions of NTP, to setting the environment variable NTPD_USE_INTERP_DANGEROUS, thus forcing the NTP time interpolation to be used. Can you elaborate more about what NTPD_USE_INTERP_DANGEROUS does? Sincerely, Ron The configuration I have is: - cable modem (with built-in router, but working as a bridge by putting my own router in a device in the DMZ). - Samknows network monitor (modified WRT54GL router) - WRT54GL router running DD-WRT firmware - Netgear 8-port consumer 1 Gb/s switch (G5108) - wired connection to ~2 year old laptop PC I only mention this to show that (a) it's not a direct connection and (b) there's no wireless involved. My aim here is simply to see what performance may be had with just Internet servers. The PC is only running NTP and monitoring software - no user programs and no interactive work, so it is a best-case scenario. Cheers, David -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
David J Taylor wrote: Terje Mathisen terje.mathisen at tmsw.no wrote in message news:is9e39-6ve2@ntp6.tmsw.no... [] You should never accept much more than 10:1 speed difference between down and up. Terje Should - I agree. But in the UK from Virgin Media I have 30 Mb/s down, 1 Mb/s up. I have been promised an upload speed increase about 18 months ago to 2 Mb/s up, which is more sensible, but that increase hasn't yet been delivered. I am also promised an increase within the next year which doubles my existing speeds, giving 60 Mb/s down and 4 Mb/s up. Not ideal, but I prefer cable to ADSL. I prefer fiber: 50/50 Mbit, very consistent ping times of a couple of ms to most no.pool.ntp.org servers. :-) Cost is about $75/month. Terje -- - Terje.Mathisen at tmsw.no almost all programming can be viewed as an exercise in caching ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/17/2012 6:00 PM, Terje Mathisen wrote: David J Taylor wrote: Terje Mathisen terje.mathisen at tmsw.no wrote in message news:is9e39-6ve2@ntp6.tmsw.no... [] You should never accept much more than 10:1 speed difference between down and up. Terje Should - I agree. But in the UK from Virgin Media I have 30 Mb/s down, 1 Mb/s up. I have been promised an upload speed increase about 18 months ago to 2 Mb/s up, which is more sensible, but that increase hasn't yet been delivered. I am also promised an increase within the next year which doubles my existing speeds, giving 60 Mb/s down and 4 Mb/s up. Not ideal, but I prefer cable to ADSL. I prefer fiber: 50/50 Mbit, very consistent ping times of a couple of ms to most no.pool.ntp.org servers. :-) Cost is about $75/month. Terje You're a lucky dude if that's even an option. For most purposes other than timing, Comcast cable does pretty good for me. Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Charles, This is a PS to my preceding reply to this. If this makes no sense, read my other reply first. Once you get your graphs working, you can select different IP's on the peerstats screen as follows. This may change in future updates to the program. Right click on the graph. An IP selection screen pops up. To select a few IP's to show: a) click the check boxes for just the IP's you want b) click OK To select all but a few IP's to show: a) click the all button b) click the check boxes to unselect the IP's you don't want c) click OK To get back to the full list of IP's a) right click to bring up the IP selection screen b) click none (Think of it as filtering none of them.) c) click OK There was something else I wanted to mention, and I'm sure I'll remember it shortly after sending this. Anyway, good luck. Sincerely, Ron On 3/16/2012 7:07 PM, Charles Elliott wrote: On the subject of accuracy, has anyone ever really looked at NTPD's offset filtering mechanism? What it does now is sort the last (about 50) offsets from smallest to largest and then prunes the smallest or largest, depending on which is further away from the average, until there are only N (I forget what N is) offset observations left. There may be at least two problems with this filtering mechanism. First, there is no apparent theory behind it; I have never seen such a crude filter that does not take into account any information inherent in the data. On the other hand, what I don't know about filters would fill all 24 volumes of an encyclopedia. Second, we know that each offset observation should have arrived about one second after the previous one, yet NTPD does not take advantage of that knowledge. There are filters, such as the Kalman filter that uses a Bayesian estimation approach to predict the next observation and adjusts it according to the prediction when it arrives, that do take advantage of previous observations. Demonstrations of the Kalman filter on the Internet show almost spectacular results. I used a Kalman filter in my clock simulation program and the results seemed pretty good. However, there are numerical analysis considerations to programming a Kalman filter as the sums and products of observations can become large in a program that runs infinitely long. Also, choosing the parameters of a Kalman filter is apparently a black art. Would it be worth it to recruit an electrical or systems engineer who claimed to know something about filtering data to take a serious look at NTPD's data filtering approach? There has to be some reason that there is a significant negative correlation between delay and offset in NTPD. There also has to be a reason that my GPS clock (BU-353, which, when it is working well, only has offset ±6 ms from zero) has a difference between about 0 and 47 ms from an NTP server on another computer that gets its time from 8 NTP stratum 2 servers over the Internet and has remarkably consistent offsets ±3.5 ms from zero. The difference between the GPS clock and the average of the stratum 2 servers appears to be a function of the time of day; it is large during the mid-part of the day, when the Internet is busy and the delay is large and quite variable between servers, and small late in the day (right now it is -0.626; 6:55 PM EST), when the delay is smaller and pretty uniform for all stratum 2 servers. Charles Elliott -Original Message- From: questions-bounces+elliott.ch=verizon@lists.ntp.org [mailto:questions-bounces+elliott.ch=verizon@lists.ntp.org] On Behalf Of Chris Albertson Sent: Thursday, March 15, 2012 5:22 PM To: unruh Cc: questions@lists.ntp.org Subject: Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error. On Thu, Mar 15, 2012 at 2:09 PM, unruhun...@invalid.ca wrote: Unfortunately it is not that simple. That rate changes by significan amounts. Thus the rate you get after a week may be very different than the rate you get after an hour. That, I submit, is the chief obstacle to having an accurate clock. And that change in rate does not fit with the Allan variance assumptions (the noise source is not of the type assumed) You are right about that. I was going to add in a bit about how to pick the best time to look at the clock tower. But left it out because the point I was making was only that these things are not NTP specific. Details after that did not contribute the the main point. Chris Albertson Redondo Beach, California -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
PPS to prior messages. I knew I'd remember this after I hit send the last time. You probably want to do any long term graphs starting right after powering off the GPS and back on. If your GPS NMEA start time is wandering, that will give you a baseline for comparison which is consistent. Also, once you've accumulated enough graph to see where the internet servers are relateive to the GPS, you can set the fudge time2 parameter to align the GPS time offsets with the average server time offsets. If you do that and restart and wait a while, the server graph lines should be almost right on top of the gps time graph. If the GPS is wandering, these server lines will appear to move away from the GPS line over time. http://dl.dropbox.com/u/9879631/peerstats%20after%20GSP%20power%20cycle%20with%20fudge%20factor%2020120316%202100.jpg Here are the server lines in my ntp.conf for the GPS. # windows lines for testing gps selected as main source - gpgga 9600 baud server 127.127.20.5 prefer minpoll 3 maxpoll 3 mode 18 fudge 127.127.20.5 time2 0.4260 refid GPS1 Note that the fudge time2 parameter required to get your GPS time comparable to internet server time will change if you change the baud rate of the GPS. Sincerely, Ron On the subject of accuracy, has anyone ever really looked at NTPD's offset filtering mechanism? What it does now is sort the last (about 50) offsets from smallest to largest and then prunes the smallest or largest, depending on which is further away from the average, until there are only N (I forget what N is) offset observations left. There may be at least two problems with this filtering mechanism. First, there is no apparent theory behind it; I have never seen such a crude filter that does not take into account any information inherent in the data. On the other hand, what I don't know about filters would fill all 24 volumes of an encyclopedia. Second, we know that each offset observation should have arrived about one second after the previous one, yet NTPD does not take advantage of that knowledge. There are filters, such as the Kalman filter that uses a Bayesian estimation approach to predict the next observation and adjusts it according to the prediction when it arrives, that do take advantage of previous observations. Demonstrations of the Kalman filter on the Internet show almost spectacular results. I used a Kalman filter in my clock simulation program and the results seemed pretty good. However, there are numerical analysis considerations to programming a Kalman filter as the sums and products of observations can become large in a program that runs infinitely long. Also, choosing the parameters of a Kalman filter is apparently a black art. Would it be worth it to recruit an electrical or systems engineer who claimed to know something about filtering data to take a serious look at NTPD's data filtering approach? There has to be some reason that there is a significant negative correlation between delay and offset in NTPD. There also has to be a reason that my GPS clock (BU-353, which, when it is working well, only has offset ±6 ms from zero) has a difference between about 0 and 47 ms from an NTP server on another computer that gets its time from 8 NTP stratum 2 servers over the Internet and has remarkably consistent offsets ±3.5 ms from zero. The difference between the GPS clock and the average of the stratum 2 servers appears to be a function of the time of day; it is large during the mid-part of the day, when the Internet is busy and the delay is large and quite variable between servers, and small late in the day (right now it is -0.626; 6:55 PM EST), when the delay is smaller and pretty uniform for all stratum 2 servers. Charles Elliott -Original Message- From: questions-bounces+elliott.ch=verizon@lists.ntp.org [mailto:questions-bounces+elliott.ch=verizon@lists.ntp.org] On Behalf Of Chris Albertson Sent: Thursday, March 15, 2012 5:22 PM To: unruh Cc: questions@lists.ntp.org Subject: Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error. On Thu, Mar 15, 2012 at 2:09 PM, unruhun...@invalid.ca wrote: Unfortunately it is not that simple. That rate changes by significan amounts. Thus the rate you get after a week may be very different than the rate you get after an hour. That, I submit, is the chief obstacle to having an accurate clock. And that change in rate does not fit with the Allan variance assumptions (the noise source is not of the type assumed) You are right about that. I was going to add in a bit about how to pick the best time to look at the clock tower. But left it out because the point I was making was only that these things are not NTP specific. Details after that did not contribute the the main point. Chris Albertson Redondo Beach, California -- (PS - If you email me and don't get a quick response, don't be concerned. I
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
unruh wrote: On 2012-03-14, David J Taylordavid-tay...@blueyonder.co.uk.invalid wrote: unruhun...@invalid.ca wrote in message news:I1T7r.36416$l12.35...@newsfe23.iad... [] Did you shut down and restart your computer? Did you perchance do this during the daylight savings time transition on a Windows system? Could the error be related to the fact that Windows like time on localtime not UTC? Windows uses UTC internally, not local time. Local time is simply a presentation layer issue. Windows is unaffected by a DST transition. That must be new, since windows certainly used to maintain system time as local time. Caused numberous headaches for people using both Windows and Linux. Windows expects the CMOS BIOS clock to be in local time (which is the opposite from Linux), but then converts this twice: First to UTC in order to be able to use the NTP protocol, and to timestamp file updates with UTC, but then it converts back to local time for almost all timekeeping functions. There is a Registry key which is supposed to allow Windows to use UTC in CMOS, but the only time I tried this (many years ago) it was somewhat broken. Terje -- - Terje.Mathisen at tmsw.no almost all programming can be viewed as an exercise in caching ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Hi David, I really appreciate all these suggestions you shared, as well as past ones. Based somewhat on been there, done that! If I decide to revamp my network, I'll probably put some of them into use. However, that's not really practical right now. All the networking gear is in the basement and all the PC's are upstairs. I've already spent 2 months working on this GPS stuff and ignoring some other things I need to be doing. I think I'm just going to focus on getting the Sure board up and running on a time server when it comes and, perhaps, use my BU-353 as a backup time source, and use the internet servers as a third level backup source. Most of the time, I won't care what they're doing Hopefully, I can spend less time thinking about time and just know that my PC's clocks are right. Regardless, I'm going to continue to have an interest in the topic and have enjoyed all the discussions and learning that have occurred. It's just that I have to do at least a few other things besides this. Sincerely, Ron Yes, sensitive those these new GPS receivers are, the basement may not be the best location! G Be certain to get the Sure antenna in a good location - it's likely sealed for outdoor use such as stuck to a car roof. I do see quite a wide variation in performance over Wi-Fi - I have two permanent PCs (Molde and Puffin) and two temporary PCs (Bergen and Mercury) connected at the moment, and averaged jitter varies a lot: 1.0 - 4.3 ms - Molde 1.1 - 2.1 ms - Puffin 1.3 - 3 ms - Bergen 4 - 12 ms - Mercury http://www.satsignal.eu/mrtg/performance_ntp.php They are all synced to a Stratum-1 FreeBSD server showing microsecond jitter, running the same NTP, with the same configuration. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f61168b.8030...@c3energy.com... [] PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. Windows is not designed as a real-time OS, and its timekeeping is not as good as other systems, but with one good local server you can sync Windows PCs to that server with something like: minpoll 5 maxpoll 5. Obviously, you would not do that over the Internet without permission. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, when the clock source is polled, say the PC clock is too fast, so NTPD slows it down. Then, when you poll the clock source again, say the PC clock is too slow, so NTPD speeds it up. Because of the varying intrinsic frequency of the clock, you can never find a clock speed that just works, because then the system goes and changes, by changes in the oscillator, how much time passes at those particular settings. It's a battle you cannot win. By polling my GPS every 8 seconds, I can keep the clock under control based on it's current needs which are varying second by second. Of course, when discussing internet servers, 30 ms of jitter doesn't help any. Windows itself doesn't vary the clock frequency unless you enable the power-saving frequency of the chipset. What I /have/ seen on Windows is poorly written third-party drivers which hold off interrupts for too long or are otherwise badly behaved, and chipsets or motherboards which are equally unfriendly. I have one application which can ruin timekeeping, given half a chance! Note the scale on PC Gemini which has an AMD processor on an ASUS A8N SLI motherboard. +/- 100 ms, not +/- 3 ms. I see large offset steps, which NTP tries its best to keep up with. [] The point is, that it's a continually moving target. The windows clock is the same way. It never runs at the same frequency from minute to minute. Even if you get it running right one minute, it's wrong the next. All PCs are a continually moving target! Some move more than others, but give a good PPS source, the major variation on even Windows PCs is due to temperature fluctuations. Let's take, for example, the TAZ computer I mentioned earlier. Forget GPS for the moment. With the default settings, NTPD will eventually be polling the internet server every 16 minutes. The problem is not exclusively that there is jitter in the time retrieved from the time server. Let's also forget that for the moment. Say we poll the internet server and it says the time is exactly 12:00:00. Rounding to the seconds level just for simplicity, say my clock says 12:00:02, so I'm 2 seconds fast. So, we slow down the clock by tweaking its parameters. Then, we wait 16 more minutes. Now the time server says 12:16:00. Say my clock says 12:15:57. Now, I'm 3 seconds slow, so we speed the clock back up. Now, theoretically, just like my pendulum clock, I should be able to get the parameters dialed in so the clock keeps time. However, behind the scenes, the intrinsic operational speed changed. While I'm sure I'm butchering the internal technicalities, let's say the clock has a speed knob, and if we set the speed knob to a value of 100, the the clock will count exactly 1 second while exactly 1 second passes. If this stayed true, NTPD would eventually set the speed knob at 100 and everybody would be happy. But the intrinsic speed of the oscillator changed, so that now setting that knob at 100 now makes the clock think 1.03 seconds has passed when, in actuality, 1 second has passed. So the clock is running fast. So, NTPD dials the speed knob back to 97. But now, the oscillator has changed again so that a setting of 97 now makes the clock think that .95 seconds have passed, when, in actuality, 1 second has passed. This is why I'm getting such wild oscillation in the graph. No matter what NTPD does to tweak the clock speed, and no matter how accurate that is at the time, that adjustment never has the same effect the next time the time server is polled. Just as setting the screw on the pendulum of my mechanical clock doesn't have the same effect tomorrow as it does today. There will never be a setting that just works. I don't think it's a game you can win on a standard windows computer. So, I'm not even inclined to play the internet NTP server game. So, by polling my GPS every 8 seconds, the computer
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/15/2012 4:34 AM, David J Taylor wrote: Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f61168b.8030...@c3energy.com... [] PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. Windows is not designed as a real-time OS, and its timekeeping is not as good as other systems, but with one good local server you can sync Windows PCs to that server with something like: minpoll 5 maxpoll 5. Obviously, you would not do that over the Internet without permission. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, when the clock source is polled, say the PC clock is too fast, so NTPD slows it down. Then, when you poll the clock source again, say the PC clock is too slow, so NTPD speeds it up. Because of the varying intrinsic frequency of the clock, you can never find a clock speed that just works, because then the system goes and changes, by changes in the oscillator, how much time passes at those particular settings. It's a battle you cannot win. By polling my GPS every 8 seconds, I can keep the clock under control based on it's current needs which are varying second by second. Of course, when discussing internet servers, 30 ms of jitter doesn't help any. Windows itself doesn't vary the clock frequency unless you enable the power-saving frequency of the chipset. What I /have/ seen on Windows is poorly written third-party drivers which hold off interrupts for too long or are otherwise badly behaved, and chipsets or motherboards which are equally unfriendly. I have one application which can ruin timekeeping, given half a chance! Note the scale on PC Gemini which has an AMD processor on an ASUS A8N SLI motherboard. +/- 100 ms, not +/- 3 ms. I see large offset steps, which NTP tries its best to keep up with. I mainly meant that operating conditions will vary the frequency of the oscillator. Speaking of which, is there a way to run ntpd and have it NOT adjust the clock at all, but still generate stats files, so I can monitor the clock frequency just based on the computer usage and nothing else? Can you elaborate on that power saving frequency thing? Is that the thing in the control panel where you set the minimum and maximum cpu frequency? [] The point is, that it's a continually moving target. The windows clock is the same way. It never runs at the same frequency from minute to minute. Even if you get it running right one minute, it's wrong the next. All PCs are a continually moving target! Some move more than others, but give a good PPS source, the major variation on even Windows PCs is due to temperature fluctuations. Let's take, for example, the TAZ computer I mentioned earlier. Forget GPS for the moment. With the default settings, NTPD will eventually be polling the internet server every 16 minutes. The problem is not exclusively that there is jitter in the time retrieved from the time server. Let's also forget that for the moment. Say we poll the internet server and it says the time is exactly 12:00:00. Rounding to the seconds level just for simplicity, say my clock says 12:00:02, so I'm 2 seconds fast. So, we slow down the clock by tweaking its parameters. Then, we wait 16 more minutes. Now the time server says 12:16:00. Say my clock says 12:15:57. Now, I'm 3 seconds slow, so we speed the clock back up. Now, theoretically, just like my pendulum clock, I should be able to get the parameters dialed in so the clock keeps time. However, behind the scenes, the intrinsic operational speed changed. While I'm sure I'm butchering the internal technicalities, let's say the clock has a speed knob, and if we set the speed knob to a value of 100, the the clock will count exactly 1 second while exactly 1 second passes. If this stayed true, NTPD would eventually set the speed knob at 100 and everybody would be happy. But the intrinsic speed of the oscillator changed, so that now setting that knob at 100 now makes the clock think 1.03 seconds has passed when, in actuality, 1 second has passed. So the clock is running fast. So, NTPD dials the speed knob back to 97. But now, the oscillator has changed again so that a setting of 97 now makes the clock think that .95 seconds have passed, when, in actuality, 1 second has passed. This is why I'm getting such wild oscillation in the graph. No matter what NTPD does to tweak the clock
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 6:42 PM, E-Mail Sent to this address will be added to the BlackLists wrote: Ron Frazier (NTP) wrote: My typical performance from internet servers is offsets of + / - 60 ms. # Why not try some that have a closer network distance, or are at least geographically closer? # Georgia / Comcast Hints: tos cohort 1 restrict source nomodify pool us.pool.ntp.org iburst # supposed to give you geographically close servers # server rolex.usg.edu iburst # server timex.usg.edu iburst # server ntp2.stsn.net iburst # server tick.gatech.edu iburst # server time.intersecur.net iburst # server nist1-atl.ustiming.org iburst # server nist1.columbiacountyga.gov iburst # _Your_ ISP router is likely already picking up the closest couple of these all by itself already # server ntp01.comcastbusiness.net iburst # server ntp02.comcastbusiness.net iburst # server ntp.whm.comcastcommercial.net iburst # server ntp01.inflow.pa.bo.comcast.net iburst # server ntp02.inflow.pa.bo.comcast.net iburst # server ntp01.cmc.co.denver.comcast.net iburst # server ntp02.cmc.co.denver.comcast.net iburst # server cr01.atlanta.ga.ibone.comcast.net iburst # server pr01.atlanta.ga.ibone.comcast.net iburst # server pe02.56marietta.ga.ibone.comcast.net iburst # server te-0-1-0-4-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-3-5-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-3-5-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-1-4-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-2-2-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-1-5-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-4-3-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-0-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-0-10-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-3-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-4-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-4-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-0-0-0-0-pe01.56marietta.ga.ibone.comcast.net iburst # ... Thanks for this info. This is a very cool server list. It's obviously relevant to my location. I'm not sure just how you did that. I'm going to keep this for reference. For the moment, I have to spend my time on getting my GPS(s) working and setting up my own time server. However, later, I might try some tinkering to improve the internet server performance. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f61e4df.4080...@c3energy.com... [] I mainly meant that operating conditions will vary the frequency of the oscillator. Speaking of which, is there a way to run ntpd and have it NOT adjust the clock at all, but still generate stats files, so I can monitor the clock frequency just based on the computer usage and nothing else? Pass. What happens if everything is noselect? Can you elaborate on that power saving frequency thing? Is that the thing in the control panel where you set the minimum and maximum cpu frequency? Likely, yes, but it varies between computers, chip makers (under different trade names), and BIOS makers. For best timekeeping, you want no clock speed variation. [] http://www.satsignal.eu/mrtg/performance_bacchus.php Wow, if I'm reading that graph right in the last link, that cpu usage spike sent your clock variation from about 100 us to 4000 us. That's amazing, and frustrating. The offset went from zero, to approximately +1.2 milliseconds. As MRTG can't plot negative numbers (it was designed for network throughput), I add 3.0 milliseconds to the reported offsets before plotting, as the left axis label says. So the range of the graph is +/- 3.0 milliseconds, plotted as 0 to 6ms. I thought all 32 bit Windows had the same kernel. You could try this defragger. I've had good luck with it, but I'm not sure which older systems it works on. http://www.auslogics.com/en/software/disk-defrag/ [] Sincerely, Ron Thanks for the suggestion, Ron, but that software, like most others, needs XP or higher. I actually use that program on one of my PCs, as it can also be called from the command-line to defragment files or directories. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-14, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: On 3/14/2012 5:04 PM, Ron Frazier (NTP) wrote: On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. I have 4 PC's connected to the LAN, plus my wife's work computer 3 days / week, and on rare occasions my son's computer. All are connected by wifi. All do pretty mundane things: web browser, email, sometimes downloading patches, sometimes doing online backup. My 4 are running NTPD. 3 run windows most of the time and dual boot into linux. My 4th machine runs linux all the time. When my wife is here, she does a remote desktop type of thing into her work system. Two of my PC's are running Vista, one is running Windows 7. Those three dual boot into Ubuntu 11.04 and the always linux machine runs Ubuntu 11.04. Almost all the discussions I've had on this mailing list are for my Windows 7 machine. The path out of my house is: PC Wifi -- Wifi router -- wired router -- cable modem -- ISP -- internet - is this PC connected over wireless or wired? Wifi G - what else is going through the router? Someone else downloading large files or using streaming audio or video? See above. Normally, no huge data hogs. - who else might be sharing your connection? It's cable. Who knows? I also get cable TV and telephone through the same wire. - what type of service do you have? Presumably not dial-up! But what speed? Comcast Cable Just tested it with speedtest.net Ping to near city: 91 ms, Download: 29.63 Mbps, Upload: 5.3 Mbps - having checked you speed, would you describe your connection as stable? See above for speed. Generally, it's very stable. However, for the purposes we're discussing, I think it's latencies and delays that are the problem. As I mentioned in my reply to David L, I'm not concerned over trying to get stellar performance from internet servers. I just want to get a good GPS server system running and use the internet servers as a backup. Sincerely, Ron Ron, Thanks for that clarification. I think you should be getting /very much/ better performance from your Internet servers. That ping is poor as well. Here I have 30 Mb/s down, just 1 Mb/s up, and NTP delays show as 18-34 ms (most in 22-30 ms). To that end, if your cable modem has multiple ports, connect one of the PCs direct to the CM and run it as your local NTP server. Wi-Fi doesn't help NTP. Later, you can add GPS/PPS to that PC as well. At the very least, connect your timekeeping PC direct to the wired router. No Wi-Fi! For my main NTP server, I got a low-powered and fan-less Intel Atom system, and it runs FreeBSD. It /only/ runs NTP, no interactive stuff at all. You might also consider getting rid of the two routers and just using the wireless one. You might also see whether you can run NTP on your router - perhaps it's a model which can run the DD-WRT firmware. I think some variants of DD-WRT can run NTP, but please check. I have a WRT 54GL in my system where I run the DD-WRT firmware. http://www.dd-wrt.com/site/index Online backup could well affect the performance of the connection for timekeeping, and your imminent Sure GPS/PPS will help enormously. Just some thoughts which may help you along the way. Cheers, David Hi David, I really appreciate all these suggestions you shared, as well as past ones. If I decide to revamp my network, I'll probably put some of them into use. However, that's not really practical right now. All the networking gear is in the basement and all the PC's are upstairs. I've already spent 2 months working on this GPS stuff and ignoring some other things I need to be doing. I think I'm just going to focus on getting the Sure board up and running on a time server when it comes and, perhaps, use my BU-353 as a backup time source, and use the internet servers as a third level backup source. Most of the time, I won't care what they're doing Hopefully, I can spend less time thinking about time and just know that my PC's clocks are right. Regardless, I'm going to continue to have an interest in the topic and have enjoyed all the discussions and learning that have occurred. It's just that I have to do at least a few other things besides this. Sincerely, Ron PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/15/2012 11:42 AM, unruh wrote: On 2012-03-14, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: On 3/14/2012 5:04 PM, Ron Frazier (NTP) wrote: On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. snip PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, Actually, the effective NTPD poll interval is abotu 8 times the stated interval. The clock filter throws away about 7 out of 8 poll results in an attempt to get rid of assmetric polls. Ie, it assumes that the shortest round trip interval out of the past 8 is the best estimate of the symmetric roundtrip and throws away the rest. Thus if you have polling every 4 min (poll interval 8) the effective interval is about every half hour. That is fine if the clock is an even half way reasonable clock (Ie rate does not change by more than say 2PPM over that time) You're saying the effective polling interval is 8x what minpoll is. However, if the access policy for a NIST server is no more than 20 times per hour or every 3 minutes, and I set minpoll to 6 or approximately every minute, even if the clock algorithm throws away 7 of 8 samples; am I not still sampling? Am I not still hitting the NIST server every minute and are they not going to ban me from accessing it if that continues? when the clock source is polled, say the PC clock is too fast, so NTPD slows it down. Then, when you poll the clock source again, say the PC clock is too slow, so NTPD speeds it up. Because of the varying intrinsic frequency of the clock, you can never find a clock speed that just works, because then the system goes and changes, by changes in the oscillator, how much time passes at those particular settings. It's a battle you cannot win. By polling my GPS every 8 seconds, I can keep the clock under control based on it's current needs which are varying second by second. Of course, when discussing internet servers, 30 ms of What are you talking about. There is no evicence either in your data or in any reports by anyone of 30ms variation is network offsets. Even on ADSL, it is in the microsecond range, not millisecond. I'm not sure exactly what you're asking. If you're referring to my comment about internet peer jitter, I occasionally see jitter numbers for internet peers on the Meinberg Time server monitor screen in the 20 - 30 ms range and more frequently see numbers in the 10 - 20 ms range for jitter. Here is a recent screen shot: http://dl.dropbox.com/u/9879631/internet%20jitter%20example.jpg Note that there are three peers with jitter in the 10 - 20 ms range. If you were asking about the offsets my computers experience using the internet as a time source, my TAZ computer polls the internet exclusively and it's offsets routinely fluctuate + / - 50 ms. http://dl.dropbox.com/u/9879631/TAZ%20loopstats%202012-03-07%20to%202012-03-14.jpg http://dl.dropbox.com/u/9879631/ntp.conf-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120313-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120314-TAZ Sincerely, Ron jitter doesn't help any. snip -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-15, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: On 3/15/2012 11:42 AM, unruh wrote: On 2012-03-14, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: On 3/14/2012 5:04 PM, Ron Frazier (NTP) wrote: On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. snip PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, Actually, the effective NTPD poll interval is abotu 8 times the stated interval. The clock filter throws away about 7 out of 8 poll results in an attempt to get rid of assmetric polls. Ie, it assumes that the shortest round trip interval out of the past 8 is the best estimate of the symmetric roundtrip and throws away the rest. Thus if you have polling every 4 min (poll interval 8) the effective interval is about every half hour. That is fine if the clock is an even half way reasonable clock (Ie rate does not change by more than say 2PPM over that time) You're saying the effective polling interval is 8x what minpoll is. However, if the access policy for a NIST server is no more than 20 times per hour or every 3 minutes, and I set minpoll to 6 or approximately every minute, even if the clock algorithm throws away 7 of 8 samples; am I not still sampling? Am I not still hitting the NIST server every minute and are they not going to ban me from accessing it if that continues? Yes. The poll interval is whatever you ( or ntp sets it to). Ie, it goes to the externam machine to get data that often. However, ntpd only uses approximately 20% of the values it gets from that remote machine. Thus as far as the ntp algorithm on your machine is concerned its effective poll interval is much longer. But why in the world are you going to the NIST servers for your time? Use the pool. A stratum 2 or 3 server can certainly give you the ms accuracy you apparently want. And if you are using gps, the only purpose of the remote servers is a) fallback, and b) getting the second right. None of those require NIST. You can cut your polling of NIST to infinity without affecting your time. when the clock source is polled, say the PC clock is too fast, so NTPD slows it down. Then, when you poll the clock source again, say the PC clock is too slow, so NTPD speeds it up. Because of the varying intrinsic frequency of the clock, you can never find a clock speed that just works, because then the system goes and changes, by changes in the oscillator, how much time passes at those particular settings. It's a battle you cannot win. By polling my GPS every 8 seconds, I can keep the clock under control based on it's current needs which are varying second by second. Of course, when discussing internet servers, 30 ms of What are you talking about. There is no evicence either in your data or in any reports by anyone of 30ms variation is network offsets. Even on ADSL, it is in the microsecond range, not millisecond. I'm not sure exactly what you're asking. If you're referring to my comment about internet peer jitter, I occasionally see jitter numbers for internet peers on the Meinberg Time server monitor screen in the 20 - 30 ms range and more frequently see numbers in the 10 - 20 ms range for jitter. Here is a recent screen shot: http://dl.dropbox.com/u/9879631/internet%20jitter%20example.jpg Note that there are three peers with jitter in the 10 - 20 ms range. Then there is something extremely wrong with those peers. Stop using them. If you were asking about the offsets my computers experience using the internet as a time source, my TAZ computer polls the internet exclusively and it's offsets routinely fluctuate + / - 50 ms. http://dl.dropbox.com/u/9879631/TAZ%20loopstats%202012-03-07%20to%202012-03-14.jpg http://dl.dropbox.com/u/9879631/ntp.conf-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120313-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120314-TAZ AGain, that is not usual and indicates something is wrong. It is not use use of the internet. Something else is very wrong. Using an internet peer 50ms away I get jitter of less than 100us. (not ms), and I think that experience is far more typical. If it is your network that is the problem then it is especially silly to use a source like NIST, since its is being totally
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-15, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: On 3/15/2012 11:42 AM, unruh wrote: On 2012-03-14, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: On 3/14/2012 5:04 PM, Ron Frazier (NTP) wrote: On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. snip PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, Actually, the effective NTPD poll interval is abotu 8 times the stated interval. The clock filter throws away about 7 out of 8 poll results in an attempt to get rid of assmetric polls. Ie, it assumes that the shortest round trip interval out of the past 8 is the best estimate of the symmetric roundtrip and throws away the rest. Thus if you have polling every 4 min (poll interval 8) the effective interval is about every half hour. That is fine if the clock is an even half way reasonable clock (Ie rate does not change by more than say 2PPM over that time) You're saying the effective polling interval is 8x what minpoll is. However, if the access policy for a NIST server is no more than 20 times per hour or every 3 minutes, and I set minpoll to 6 or approximately every minute, even if the clock algorithm throws away 7 of 8 samples; am I not still sampling? Am I not still hitting the NIST server every minute and are they not going to ban me from accessing it if that continues? Yes. The poll interval is whatever you ( or ntp sets it to). Ie, it goes to the externam machine to get data that often. However, ntpd only uses approximately 20% of the values it gets from that remote machine. Thus as far as the ntp algorithm on your machine is concerned its effective poll interval is much longer. But why in the world are you going to the NIST servers for your time? Use the pool. A stratum 2 or 3 server can certainly give you the ms accuracy you apparently want. And if you are using gps, the only purpose of the remote servers is a) fallback, and b) getting the second right. None of those require NIST. You can cut your polling of NIST to infinity without affecting your time. Allow me to explain. A long time ago in a galaxy far far away, I wanted to synchronize the clocks of my Windows machines. This was before I knew anything about Linux or NTP. However, being a US citizen, I knew our government provided the National Institute of Standards and Technology (NIST) agency, and that they had a Time and Frequency division, just to help handle such things for US citizens and the world. So, I went here to the Time and Frequency division (the link address may have changed over time): http://www.nist.gov/pml/div688/ And then here to the Internet Time Service: http://www.nist.gov/pml/div688/grp40/its.cfm And I downloaed their little Windows time program here: http://www.nist.gov/pml/div688/grp40/upload/nistime-32bit.exe This thing is designed to poll the NIST servers (and only the NIST servers) up to every 4 hours and set your clock. That's all it does, and I don't think it tries to tinker with the clock frequency. I ran that for many years, and still, sometimes, would get a couple of seconds drift between polling intervals. There is nothing on the website to discourage average Joe users from using the system. A couple of years ago, I learned about and started using Ubuntu. For timekeeping there, I used NTP. At the time, there was a graphical interface to it and I picked about 8 random servers by ticking the check boxes and let it run. Back around Christmas 2011, I got some atomic wall clocks and an atomic watch and got fascinated with the idea of getting more accuracy for the PC's. That led me to doing reserach and that led me to install the Meinberg port of NTP. Of course, then, I had to figure out how to add servers and set up ntp.conf. In answer to your question, I'm using NIST because: a) My government provides it and, through their website, encourages the public to use it. b) I was familiar with it. c) I thought it would be more accurate. Having said that, I could discontinue using it if the need arises. So, I put 4 NIST servers, 4 stratum 2 servers, 4 US pool servers, and just for good measure, the Ubuntu time server in my ntp.conf. Since all queries exiting my house show the same
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Basically you have a trade-off between frequency error and time offset. IIRC, polling quickly will reduce the offset, but you get greater frequency variations. There is nothing magic about NTP. You have the exact same trade off if you try and sync your wrist watch to the big town clock on the tower. If you look at the tower every hour and set the wrist watch to match you will have very small offset between the watch and the big clock. But if you do this the _rate_ of your wrist watch will never be spot-on except by luck. On the other hand if you sync your watch once then wait a week or a month to see how much it has gained or lost you may see quite a bit of offset but now you have good data on the relative rates of the two clocks NTP did not invent this trade off, it existed 200 years ago. Chris Albertson Redondo Beach, California ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-15, Chris Albertson albertson.ch...@gmail.com wrote: Basically you have a trade-off between frequency error and time offset. IIRC, polling quickly will reduce the offset, but you get greater frequency variations. There is nothing magic about NTP. You have the exact same trade off if you try and sync your wrist watch to the big town clock on the tower. If you look at the tower every hour and set the wrist watch to match you will have very small offset between the watch and the big clock. But if you do this the _rate_ of your wrist watch will never be spot-on except by luck. On the other hand if you sync your watch once then wait a week or a month to see how much it has gained or lost you may see quite a bit of offset but now you have good data on the relative rates of the two clocks NTP did not invent this trade off, it existed 200 years ago. Unfortunately it is not that simple. That rate changes by significan amounts. Thus the rate you get after a week may be very different than the rate you get after an hour. That, I submit, is the chief obstacle to having an accurate clock. And that change in rate does not fit with the Allan variance assumptions (the noise source is not of the type assumed) Chris Albertson Redondo Beach, California ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-15, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: On 2012-03-15, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: On 3/15/2012 11:42 AM, unruh wrote: On 2012-03-14, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: On 3/14/2012 5:04 PM, Ron Frazier (NTP) wrote: On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. snip PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, Actually, the effective NTPD poll interval is abotu 8 times the stated interval. The clock filter throws away about 7 out of 8 poll results in an attempt to get rid of assmetric polls. Ie, it assumes that the shortest round trip interval out of the past 8 is the best estimate of the symmetric roundtrip and throws away the rest. Thus if you have polling every 4 min (poll interval 8) the effective interval is about every half hour. That is fine if the clock is an even half way reasonable clock (Ie rate does not change by more than say 2PPM over that time) You're saying the effective polling interval is 8x what minpoll is. However, if the access policy for a NIST server is no more than 20 times per hour or every 3 minutes, and I set minpoll to 6 or approximately every minute, even if the clock algorithm throws away 7 of 8 samples; am I not still sampling? Am I not still hitting the NIST server every minute and are they not going to ban me from accessing it if that continues? Yes. The poll interval is whatever you ( or ntp sets it to). Ie, it goes to the externam machine to get data that often. However, ntpd only uses approximately 20% of the values it gets from that remote machine. Thus as far as the ntp algorithm on your machine is concerned its effective poll interval is much longer. But why in the world are you going to the NIST servers for your time? Use the pool. A stratum 2 or 3 server can certainly give you the ms accuracy you apparently want. And if you are using gps, the only purpose of the remote servers is a) fallback, and b) getting the second right. None of those require NIST. You can cut your polling of NIST to infinity without affecting your time. Allow me to explain. A long time ago in a galaxy far far away, I wanted to synchronize the clocks of my Windows machines. This was before I knew anything about Linux or NTP. However, being a US citizen, I knew our government provided the National Institute of Standards and Technology (NIST) agency, and that they had a Time and Frequency division, just to help handle such things for US citizens and the world. So, I went here to the Time and Frequency division (the link address may have changed over time): http://www.nist.gov/pml/div688/ And then here to the Internet Time Service: http://www.nist.gov/pml/div688/grp40/its.cfm And I downloaed their little Windows time program here: http://www.nist.gov/pml/div688/grp40/upload/nistime-32bit.exe This thing is designed to poll the NIST servers (and only the NIST servers) up to every 4 hours and set your clock. That's all it does, and I don't think it tries to tinker with the clock frequency. I ran that for many years, and still, sometimes, would get a couple of seconds drift between polling intervals. There is nothing on the website to discourage average Joe users from using the system. A couple of years ago, I learned about and started using Ubuntu. For timekeeping there, I used NTP. At the time, there was a graphical interface to it and I picked about 8 random servers by ticking the check boxes and let it run. Back around Christmas 2011, I got some atomic wall clocks and an atomic watch and got fascinated with the idea of getting more accuracy for the PC's. That led me to doing reserach and that led me to install the Meinberg port of NTP. Of course, then, I had to figure out how to add servers and set up ntp.conf. In answer to your question, I'm using NIST because: a) My government provides it and, through their website, encourages the public to use it. b) I was familiar with it. c) I thought it would be more accurate. Having said that, I could discontinue using it if the need arises. And since your early forays into
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On Thu, Mar 15, 2012 at 2:09 PM, unruh un...@invalid.ca wrote: Unfortunately it is not that simple. That rate changes by significan amounts. Thus the rate you get after a week may be very different than the rate you get after an hour. That, I submit, is the chief obstacle to having an accurate clock. And that change in rate does not fit with the Allan variance assumptions (the noise source is not of the type assumed) You are right about that. I was going to add in a bit about how to pick the best time to look at the clock tower. But left it out because the point I was making was only that these things are not NTP specific. Details after that did not contribute the the main point. Chris Albertson Redondo Beach, California ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On Mar 15, 2012, at 2:06 PM, unruh wrote: Before hitting on 13 different servers, it would be better tofigure out what the problem is that is giving you such lousy results. It's not hard to make an educated guess as to the cause; since the topology is: PC Wifi -- Wifi router -- wired router -- cable modem -- ISP -- internet WiFi radios are half-duplex since only one radio can be sending at a time, else you get a backoff similar to ethernet collisions on a hub. If the WiFi link is idle (and nobody else is using 2.4 GHz bands nearby!), one ought to be able to get reasonably good timing, but as soon as there is any significant congestion from other machines (or phones, tablets, etc), a wireless NTP client will experience heavy jitter on the level of tens of milliseconds. [ That doesn't get Ron up to 120ms, but if double-NAT is going on, or the routers aren't playing nice with each other for some other reason, it wouldn't be surprising to end up with marginal timekeeping results. Plug something directly into ethernet jack on the wired router and retry ] Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Hi David, OK, you've convinced me. I've put noselect on the GPS and taken it away on the New York NIST server. The NY NIST server is now preferred and polling from 1 - 4 minutes. Hopefully they won't ban me for hitting it too often. I think the NIST server will have more jitter, particularly as the polling interval increases, but we'll see what happens. Sincerely, Ron It will be interesting to see what difference that makes, Ron, and I look forward to seeing the graphs. Personally, I saw little to choose between the various Internet servers you had, and I would have allowed four or five to be selected rather than just one, as that would not then leave NTP open to failure of the one server, and it would allow NTP to select what It believes to be the best. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
unruh un...@invalid.ca wrote in message news:I1T7r.36416$l12.35...@newsfe23.iad... [] Did you shut down and restart your computer? Did you perchance do this during the daylight savings time transition on a Windows system? Could the error be related to the fact that Windows like time on localtime not UTC? Windows uses UTC internally, not local time. Local time is simply a presentation layer issue. Windows is unaffected by a DST transition. David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-13, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looking at this graph, I see that the nmea source was already difting before the sudden jump. It lost 30ms wrt the other servers in the 20 hours beforehand. Then it went crazy for a while and jumped to 80 ms ahead. I agree that this does seem to be that gps device. Which one is it? But that jump is 120ms not 50 sec. I recently had a Garmin 18 go nuts-- giving massive amounts of noise. Here is what was showing on the Meinberg Time Server Monitor when I woke up: http://dl.dropbox.com/u/9879631/drifting02a%20-%20insane.jpg And the graph of the peerstats for that time: http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg The clock error was REAL, as confirmed by my atomic wrist watch. However, the loopstats graph for the same time period shows no problem with the GPS: http://dl.dropbox.com/u/9879631/drifting02c%20-%20gps%20NOT%20insane.jpg So, I shut down NTPD and reset the time with a batch file that calls ntpdate and querys the New York NIST server: http://dl.dropbox.com/u/9879631/drifting02d%20-%20set%20with%20ntpdate.jpg Here is the time server monitor shortly after NTPD restart: http://dl.dropbox.com/u/9879631/drifting02e%20-%20shortly%20after%20ntpd%20restart.jpg And after a 2nd restart: http://dl.dropbox.com/u/9879631/drifting02f%20-%202nd%20ntpd%20restart%20after%20insane.jpg And here are the current peerstats, which look normal. The offset to the internet servers tends to drift and will eventually cross the zero line and get positive. http://dl.dropbox.com/u/9879631/drifting02g%20-%20peerstats%20after%20insane%20and%202nd%20reset.jpg The GPS appears to have been stable all through this, and was never powered off or unplugged. It looks like NTPD went crazy and reset my clock for some reason. It also reset all of the remote servers at the same time? Since it was an offset of your system with respect to the remote systems. Here are the peerstats and loopstats during the insane period. http://dl.dropbox.com/u/9879631/loopstats.20120313-1-restart%20around%201350%20utc http://dl.dropbox.com/u/9879631/peerstats.20120313-1-insane Here is my current ntp.conf: http://dl.dropbox.com/u/9879631/ntp.conf My system is Windows 7, NTP 4.2.7p259, GPS GlobalSat BU-353 USB NMEA only with GPGGA sentence at 57,600 baud. If anyone can shed some light on what happened, please do. It's driving me bonkers. I don't believe the GPS is at fault, and I suspect NTPD. Again, the remote servers all agree. The GPS time does not (driving your system time) . Thanks in advance. Sincerely, Ron ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) wrote: On 3/13/2012 5:39 PM, David Lord wrote: Ron Frazier (NTP) wrote: On 3/13/2012 2:40 PM, Chris Albertson wrote: On Tue, Mar 13, 2012 at 9:52 AM, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looks to me like that GPS, maybe lock lock and was able to coast on hold over for a while then fell off a cliff. If you have a log of satellite signal to noise ratios you might be able to figure out why. The internet servers appear to be 100% reliable seeing as they all agree. These kinds of things are why some hobbysts end up buying multiple GPS (different brands) Otherwise it is hard to sort out a GPS firmware bug from a solar storm or just that there were not sats visable to your indoor antenna for a few minutes I think your goal is to learn about all of this so these problems are a good thing. No one learns much from working systems. But if the goal is a reliable NTP server, the pool NTP servers can't be beat except by a good timing mode GPS, that has good self diagnostics and PPS. The self diagnostics part is important The link you quoted just now is not the one that bugs me so much today. This one is the one that bugs me today, where my clock was stepped by 50 seconds for some reason. http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg I'm on a text mostly system so why do you require me to use a web browser to help you? What is your current ntp.conf, ntpd version, operating system and response of ntpq -crv -p after ntpd has had at least a day to stabilise? Not so long ago you posted that you had an offset within +/- 6 ms but didn't give any evidence. Response from ntpq -crv -p or better a day from peer_summary or couple of days from loop_summary would be enough. David When I get the Sure board, I plan to hang it on the pc in tandem with this GPS and compare. Sincerely, Ron Hi David L., I just figured the images and dropbox would be the fastest way to distribute the information. I didn't think I could attach files to messages going to the NTP questions mailing list. I'll be glad to provide stats files if anyone is interested. Just tell me how to distribute them. Based on advice from David Taylor, I just changed my configuration. Previously, I had only the GPS selectable and the internet servers noselected for testing purposes to try to determine where a slow drifting behavior is coming from. Now, I have changed that so the GPS is noselected and the New York NIST server is the preferred and only selectable peer. I am monitoring the GPS and other internet servers for comparison. So, it will be a couple of days before this configuration accumulates some stats files. Even though I have been tinkering with this GPS for a couple of months, I have never seen anything like the 50 second jump I started this thread about. That problem may not be reproducible for some time, if at all. Faulty GPS, incorrect configuration, anybodies guess? Internet servers give me an rms offset about 610us, whilst GPS with PPS gives about 4us. There are day to day variations and some rare 'events' when GPS or an internet server goes bad. GPS without PPS, Globalsat BR304, wasn't worth using as ntp source due to large variations in offset from the NMEA sentences that were tried with RMC being best giving 50% of offsets under 10ms but maximum offsets being near 100ms. MSF radioclock mostly has offset below 1500us. DCF radioclock can have offset below 100us but reception is too variable being non existant for some periods each day. I use mrtg to monitor my servers because it's easy to see when a fault occurs but the graphs don't help to locate the cause. (1) Internet servers only, rms offset 609us: loopstats.20120313 loop 71, 368+/-1360.5, rms 608.7, freq 1.24+/-0.173, var 0.112 peerstats.20120313 ident cnt mean rms max delay dist disp == 81.187.61.74 70 -0.1230.5061.2550.416 21.959 10.600 158.152.1.76 690.1180.9272.025 17.864 28.174 10.205 130.88.200.4 706.0991.6863.553 23.694 33.833 10.345 90.155.53.94 68 -0.0341.1272.952 17.376 29.720 10.938 79.135.97.79 710.5730.7221.388 31.522 38.183 10.618 130.159.196.117 750.3130.9671.830 27.958 35.9169.516 (2) GPS with PPS, local and internet servers, rms offset 3.4us: loopstats.20120313 loop 1350, 4+/-25.6, rms 3.4, freq -35.28+/-0.222, var 0.067 peerstats.20120313 ident cnt mean rms max delay
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 2:46 AM, David J Taylor wrote: Hi David, OK, you've convinced me. I've put noselect on the GPS and taken it away on the New York NIST server. The NY NIST server is now preferred and polling from 1 - 4 minutes. Hopefully they won't ban me for hitting it too often. I think the NIST server will have more jitter, particularly as the polling interval increases, but we'll see what happens. Sincerely, Ron It will be interesting to see what difference that makes, Ron, and I look forward to seeing the graphs. Personally, I saw little to choose between the various Internet servers you had, and I would have allowed four or five to be selected rather than just one, as that would not then leave NTP open to failure of the one server, and it would allow NTP to select what It believes to be the best. Cheers, David Hi David, OK. You asked for it. 8-) When I'm through testing, I'll open up the other internet servers as a backup in case the GPS fails. For now, I'm just running with one clock source at a time. Still trying to document and chase down this wandering effect. I ran with NY NIST as the only selectable clock source and monitoring the GPS for comparison all night. The results were horrible. My offsets from NIST time were in the + 65 ms / - 75 ms range. I had the polling interval set to start at 1 minute and go up to 4 minutes. There is way too much clock wander to even think about testing the accuracy of the GPS. I've gone back to polling the GPS every 8 seconds as the sole selectable clock source and monitoring the internet servers for comparison. Over the short term, minutes to hours, my GPS, even with NMEA only, is by far the most accurate time source I have. Even if the NMEA signal wanders 70 ms either way over the course of a few days, it won't get any further off than I did using the internet server, and the clock will be much more consistent over shorter time frames. Here are the graphs. http://dl.dropbox.com/u/9879631/nynist01.jpg http://dl.dropbox.com/u/9879631/nynist02.jpg Sincerely, Ron PS - By the way, I've changed my ntp.conf a few times since I originally started this thread, including the one in dropbox. What's in there now is my current ntp.conf, not the one that was in there when I started the thread. -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 3:03 AM, unruh wrote: On 2012-03-13, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looking at this graph, I see that the nmea source was already difting before the sudden jump. It lost 30ms wrt the other servers in the 20 hours beforehand. Then it went crazy for a while and jumped to 80 ms ahead. I agree that this does seem to be that gps device. Which one is it? But that jump is 120ms not 50 sec. I recently had a Garmin 18 go nuts-- giving massive amounts of noise. That's very interesting. David Taylor also said he saw this NMEA wandering effect on the Garmin. Did your Garmin recover? And, is it based on a SIRF chipset? In another thread, someone else with a BU-353 said he saw an offset storm like the one in my graph. Sincerely, Ron Here is what was showing on the Meinberg Time Server Monitor when I woke up: http://dl.dropbox.com/u/9879631/drifting02a%20-%20insane.jpg And the graph of the peerstats for that time: http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg The clock error was REAL, as confirmed by my atomic wrist watch. However, the loopstats graph for the same time period shows no problem with the GPS: http://dl.dropbox.com/u/9879631/drifting02c%20-%20gps%20NOT%20insane.jpg So, I shut down NTPD and reset the time with a batch file that calls ntpdate and querys the New York NIST server: http://dl.dropbox.com/u/9879631/drifting02d%20-%20set%20with%20ntpdate.jpg Here is the time server monitor shortly after NTPD restart: http://dl.dropbox.com/u/9879631/drifting02e%20-%20shortly%20after%20ntpd%20restart.jpg And after a 2nd restart: http://dl.dropbox.com/u/9879631/drifting02f%20-%202nd%20ntpd%20restart%20after%20insane.jpg And here are the current peerstats, which look normal. The offset to the internet servers tends to drift and will eventually cross the zero line and get positive. http://dl.dropbox.com/u/9879631/drifting02g%20-%20peerstats%20after%20insane%20and%202nd%20reset.jpg The GPS appears to have been stable all through this, and was never powered off or unplugged. It looks like NTPD went crazy and reset my clock for some reason. It also reset all of the remote servers at the same time? Since it was an offset of your system with respect to the remote systems. Here are the peerstats and loopstats during the insane period. http://dl.dropbox.com/u/9879631/loopstats.20120313-1-restart%20around%201350%20utc http://dl.dropbox.com/u/9879631/peerstats.20120313-1-insane Here is my current ntp.conf: http://dl.dropbox.com/u/9879631/ntp.conf My system is Windows 7, NTP 4.2.7p259, GPS GlobalSat BU-353 USB NMEA only with GPGGA sentence at 57,600 baud. If anyone can shed some light on what happened, please do. It's driving me bonkers. I don't believe the GPS is at fault, and I suspect NTPD. Again, the remote servers all agree. The GPS time does not (driving your system time) . Thanks in advance. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 7:14 AM, David Lord wrote: Ron Frazier (NTP) wrote: On 3/13/2012 5:39 PM, David Lord wrote: Ron Frazier (NTP) wrote: On 3/13/2012 2:40 PM, Chris Albertson wrote: On Tue, Mar 13, 2012 at 9:52 AM, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looks to me like that GPS, maybe lock lock and was able to coast on hold over for a while then fell off a cliff. If you have a log of satellite signal to noise ratios you might be able to figure out why. The internet servers appear to be 100% reliable seeing as they all agree. These kinds of things are why some hobbysts end up buying multiple GPS (different brands) Otherwise it is hard to sort out a GPS firmware bug from a solar storm or just that there were not sats visable to your indoor antenna for a few minutes I think your goal is to learn about all of this so these problems are a good thing. No one learns much from working systems. But if the goal is a reliable NTP server, the pool NTP servers can't be beat except by a good timing mode GPS, that has good self diagnostics and PPS. The self diagnostics part is important The link you quoted just now is not the one that bugs me so much today. This one is the one that bugs me today, where my clock was stepped by 50 seconds for some reason. http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg I'm on a text mostly system so why do you require me to use a web browser to help you? What is your current ntp.conf, ntpd version, operating system and response of ntpq -crv -p after ntpd has had at least a day to stabilise? Not so long ago you posted that you had an offset within +/- 6 ms but didn't give any evidence. Response from ntpq -crv -p or better a day from peer_summary or couple of days from loop_summary would be enough. David When I get the Sure board, I plan to hang it on the pc in tandem with this GPS and compare. Sincerely, Ron Hi David L., I just figured the images and dropbox would be the fastest way to distribute the information. I didn't think I could attach files to messages going to the NTP questions mailing list. I'll be glad to provide stats files if anyone is interested. Just tell me how to distribute them. Based on advice from David Taylor, I just changed my configuration. Previously, I had only the GPS selectable and the internet servers noselected for testing purposes to try to determine where a slow drifting behavior is coming from. Now, I have changed that so the GPS is noselected and the New York NIST server is the preferred and only selectable peer. I am monitoring the GPS and other internet servers for comparison. So, it will be a couple of days before this configuration accumulates some stats files. Even though I have been tinkering with this GPS for a couple of months, I have never seen anything like the 50 second jump I started this thread about. That problem may not be reproducible for some time, if at all. Faulty GPS, incorrect configuration, anybodies guess? Internet servers give me an rms offset about 610us, whilst GPS with PPS gives about 4us. There are day to day variations and some rare 'events' when GPS or an internet server goes bad. You must LIVE on the internet backbone and have blazing fast routers to get performance like that. My typical performance from internet servers is offsets of + / - 60 ms. GPS without PPS, Globalsat BR304, wasn't worth using as ntp source due to large variations in offset from the NMEA sentences that were tried with RMC being best giving 50% of offsets under 10ms but maximum offsets being near 100ms. With my BU-353, which is similar, by setting the baud rate to 57,600, programming for ONLY GPGGA sentence, and polling every 8 seconds, I can keep my offsets from the GPS's estimate of true time to usually around + / - 5 ms with spikes to 10 ms and almost never higher. If I use ONLY the GPZDA sentence, which is essentially fixed length and reports only time, I can get that down to + / - 3 ms most of the time with spikes to 6 ms. However, David Taylor pointed out that the GPZDA sentence doesn't have any validity check field, and the GPGGA sentence does. We verified that the refclock.c code does check for this. So, I'm back to using GPGGA even with a bit more jitter. That way, if the GPS fails, ntpd is more likely to react gracefully. The problem with using my BU-353 in this way is that the start times for the NMEA sentence seem to wander over about 60 ms in either direction over a period of about 4 days. So, over a few days, my computer's time will drift away from true UTC time by that amount, and back
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Hi David, OK. You asked for it. 8-) Well, I actually suggested /all/ the Internet servers being enabled, allowing NTP to make its best choice. When I'm through testing, I'll open up the other internet servers as a backup in case the GPS fails. For now, I'm just running with one clock source at a time. Still trying to document and chase down this wandering effect. I ran with NY NIST as the only selectable clock source and monitoring the GPS for comparison all night. The results were horrible. My offsets from NIST time were in the + 65 ms / - 75 ms range. I had the polling interval set to start at 1 minute and go up to 4 minutes. There is way too much clock wander to even think about testing the accuracy of the GPS. I've gone back to polling the GPS every 8 seconds as the sole selectable clock source and monitoring the internet servers for comparison. Over the short term, minutes to hours, my GPS, even with NMEA only, is by far the most accurate time source I have. Even if the NMEA signal wanders 70 ms either way over the course of a few days, it won't get any further off than I did using the internet server, and the clock will be much more consistent over shorter time frames. Here are the graphs. http://dl.dropbox.com/u/9879631/nynist01.jpg http://dl.dropbox.com/u/9879631/nynist02.jpg Sincerely, Ron I'm not surprised that using a single Internet server is worse than the GPS/USB, but that's not how NTP is designed to work. With two or more Internet servers active, your GPS 50-second glitch would not have affected your PC's timekeeping anything like as severely, when you have the GPS/USB included to help improve the offset and more like the narrow band (about 15 milliseconds wide) shown in: http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 10:41 AM, David J Taylor wrote: Hi David, OK. You asked for it. 8-) Well, I actually suggested /all/ the Internet servers being enabled, allowing NTP to make its best choice. When I'm through testing, I'll open up the other internet servers as a backup in case the GPS fails. For now, I'm just running with one clock source at a time. Still trying to document and chase down this wandering effect. I ran with NY NIST as the only selectable clock source and monitoring the GPS for comparison all night. The results were horrible. My offsets from NIST time were in the + 65 ms / - 75 ms range. I had the polling interval set to start at 1 minute and go up to 4 minutes. There is way too much clock wander to even think about testing the accuracy of the GPS. I've gone back to polling the GPS every 8 seconds as the sole selectable clock source and monitoring the internet servers for comparison. Over the short term, minutes to hours, my GPS, even with NMEA only, is by far the most accurate time source I have. Even if the NMEA signal wanders 70 ms either way over the course of a few days, it won't get any further off than I did using the internet server, and the clock will be much more consistent over shorter time frames. Here are the graphs. http://dl.dropbox.com/u/9879631/nynist01.jpg http://dl.dropbox.com/u/9879631/nynist02.jpg Sincerely, Ron I'm not surprised that using a single Internet server is worse than the GPS/USB, but that's not how NTP is designed to work. With two or more Internet servers active, your GPS 50-second glitch would not have affected your PC's timekeeping anything like as severely, when you have the GPS/USB included to help improve the offset and more like the narrow band (about 15 milliseconds wide) shown in: http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Cheers, David OK. Here are the loopstats from another computer for 7 days (in the chart). I don't have any peerstats for it. It has the same server list. One is preferred. All servers are active. Performance is no better. http://dl.dropbox.com/u/9879631/TAZ%20loopstats%202012-03-07%20to%202012-03-14.jpg http://dl.dropbox.com/u/9879631/ntp.conf-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120313-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120314-TAZ Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) wrote: On 3/14/2012 7:14 AM, David Lord wrote: Ron Frazier (NTP) wrote: Based on advice from David Taylor, I just changed my configuration. Previously, I had only the GPS selectable and the internet servers noselected for testing purposes to try to determine where a slow drifting behavior is coming from. Now, I have changed that so the GPS is noselected and the New York NIST server is the preferred and only selectable peer. I am monitoring the GPS and other internet servers for comparison. So, it will be a couple of days before this configuration accumulates some stats files. Even though I have been tinkering with this GPS for a couple of months, I have never seen anything like the 50 second jump I started this thread about. That problem may not be reproducible for some time, if at all. Faulty GPS, incorrect configuration, anybodies guess? Internet servers give me an rms offset about 610us, whilst GPS with PPS gives about 4us. There are day to day variations and some rare 'events' when GPS or an internet server goes bad. You must LIVE on the internet backbone and have blazing fast routers to get performance like that. My typical performance from internet servers is offsets of + / - 60 ms. Hi Ron I don't live on the internet backbone nor do I have blazingly fast anything. I have a slow 2Mbit/s ADSL connection. You must have a very broken internet connection to get offsets that high. A few times I've been routed by satelite or some wireless connection and I also sometimes use mobile broadband but still don't see offsets as large as 60ms. GPS without PPS, Globalsat BR304, wasn't worth using as ntp source due to large variations in offset from the NMEA sentences that were tried with RMC being best giving 50% of offsets under 10ms but maximum offsets being near 100ms. With my BU-353, which is similar, by setting the baud rate to 57,600, programming for ONLY GPGGA sentence, and polling every 8 seconds, I can keep my offsets from the GPS's estimate of true time to usually around + / - 5 ms with spikes to 10 ms and almost never higher. If I use ONLY the GPZDA sentence, which is essentially fixed length and reports only time, I can get that down to + / - 3 ms most of the time with spikes to 6 ms. However, David Taylor pointed out that the GPZDA sentence doesn't have any validity check field, and the GPGGA sentence does. We verified that the refclock.c code does check for this. So, I'm back to using GPGGA even with a bit more jitter. That way, if the GPS fails, ntpd is more likely to react gracefully. What are you using as a reference to verify the offsets you get are offsets from UTC? The problem with using my BU-353 in this way is that the start times for the NMEA sentence seem to wander over about 60 ms in either direction over a period of about 4 days. So, over a few days, my computer's time will drift away from true UTC time by that amount, and back again. However, over the short term, my pc's clock is much more stable than if I use internet servers as my primary source. Can you give your ntp.conf that result in that level of offset from internet servers? The ntp distribution might still have some advice and tools for selection of suitable sources. I just select from the list of uk public servers at ntp.org and try each one to get rtt and their sources then select the ones that are closest with different sources. If you don't want to do that you can specify your local ntp pool. You should select at least four sources. I have about eight different sources split between ntp0.lordynet.org and ntp1.lordynet.org. David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f60bc0c.8040...@c3energy.com... [] OK. Here are the loopstats from another computer for 7 days (in the chart). I don't have any peerstats for it. It has the same server list. One is preferred. All servers are active. Performance is no better. http://dl.dropbox.com/u/9879631/TAZ%20loopstats%202012-03-07%20to%202012-03-14.jpg http://dl.dropbox.com/u/9879631/ntp.conf-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120313-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120314-TAZ Sincerely, Ron That's horrible, Ron! Much worse than David Lord reports. It makes me want to ask a number of questions: - is this PC connected over wireless or wired? - what else is going through the router? Someone else downloading large files or using streaming audio or video? - who else might be sharing your connection? - what type of service do you have? Presumably not dial-up! But what speed? - having checked you speed, would you describe your connection as stable? The quiet periods (e.g early Monday morning, Tuesday lunchtime) are much nearer to what I would hope for, which makes me wonder whether something is interfering with the connection outside those times. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-14, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: unruh un...@invalid.ca wrote in message news:I1T7r.36416$l12.35...@newsfe23.iad... [] Did you shut down and restart your computer? Did you perchance do this during the daylight savings time transition on a Windows system? Could the error be related to the fact that Windows like time on localtime not UTC? Windows uses UTC internally, not local time. Local time is simply a presentation layer issue. Windows is unaffected by a DST transition. That must be new, since windows certainly used to maintain system time as local time. Caused numberous headaches for people using both Windows and Linux. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-14, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: On 3/14/2012 3:03 AM, unruh wrote: On 2012-03-13, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looking at this graph, I see that the nmea source was already difting before the sudden jump. It lost 30ms wrt the other servers in the 20 hours beforehand. Then it went crazy for a while and jumped to 80 ms ahead. I agree that this does seem to be that gps device. Which one is it? But that jump is 120ms not 50 sec. I recently had a Garmin 18 go nuts-- giving massive amounts of noise. That's very interesting. David Taylor also said he saw this NMEA wandering effect on the Garmin. Did your Garmin recover? And, is it based on a SIRF chipset? No idea. It is the old Garmin 18LVC (not 18x) In another thread, someone else with a BU-353 said he saw an offset storm like the one in my graph. Sincerely, Ron Here is what was showing on the Meinberg Time Server Monitor when I woke up: http://dl.dropbox.com/u/9879631/drifting02a%20-%20insane.jpg And the graph of the peerstats for that time: http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg The clock error was REAL, as confirmed by my atomic wrist watch. However, the loopstats graph for the same time period shows no problem with the GPS: http://dl.dropbox.com/u/9879631/drifting02c%20-%20gps%20NOT%20insane.jpg So, I shut down NTPD and reset the time with a batch file that calls ntpdate and querys the New York NIST server: http://dl.dropbox.com/u/9879631/drifting02d%20-%20set%20with%20ntpdate.jpg Here is the time server monitor shortly after NTPD restart: http://dl.dropbox.com/u/9879631/drifting02e%20-%20shortly%20after%20ntpd%20restart.jpg And after a 2nd restart: http://dl.dropbox.com/u/9879631/drifting02f%20-%202nd%20ntpd%20restart%20after%20insane.jpg And here are the current peerstats, which look normal. The offset to the internet servers tends to drift and will eventually cross the zero line and get positive. http://dl.dropbox.com/u/9879631/drifting02g%20-%20peerstats%20after%20insane%20and%202nd%20reset.jpg The GPS appears to have been stable all through this, and was never powered off or unplugged. It looks like NTPD went crazy and reset my clock for some reason. It also reset all of the remote servers at the same time? Since it was an offset of your system with respect to the remote systems. Here are the peerstats and loopstats during the insane period. http://dl.dropbox.com/u/9879631/loopstats.20120313-1-restart%20around%201350%20utc http://dl.dropbox.com/u/9879631/peerstats.20120313-1-insane Here is my current ntp.conf: http://dl.dropbox.com/u/9879631/ntp.conf My system is Windows 7, NTP 4.2.7p259, GPS GlobalSat BU-353 USB NMEA only with GPGGA sentence at 57,600 baud. If anyone can shed some light on what happened, please do. It's driving me bonkers. I don't believe the GPS is at fault, and I suspect NTPD. Again, the remote servers all agree. The GPS time does not (driving your system time) . Thanks in advance. Sincerely, Ron ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
unruh un...@invalid.ca wrote in message news:HI48r.39505$zd5.1...@newsfe12.iad... On 2012-03-14, David J Taylor wrote: [] Windows uses UTC internally, not local time. Local time is simply a presentation layer issue. Windows is unaffected by a DST transition. That must be new, since windows certainly used to maintain system time as local time. Caused numberous headaches for people using both Windows and Linux. Not new at all. It's been that way since 1992 for the whole of the NT family. Perhaps you are thinking of what is stored in the real-time clock chip? Windows and UNIX have different conventions for that, Windows using wall-clock time. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 12:35 PM, David J Taylor wrote: Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f60bc0c.8040...@c3energy.com... [] OK. Here are the loopstats from another computer for 7 days (in the chart). I don't have any peerstats for it. It has the same server list. One is preferred. All servers are active. Performance is no better. http://dl.dropbox.com/u/9879631/TAZ%20loopstats%202012-03-07%20to%202012-03-14.jpg http://dl.dropbox.com/u/9879631/ntp.conf-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120313-TAZ http://dl.dropbox.com/u/9879631/loopstats.20120314-TAZ Sincerely, Ron That's horrible, Ron! Much worse than David Lord reports. It makes me want to ask a number of questions: Hi David T, NOW you understand. I have 4 PC's connected to the LAN, plus my wife's work computer 3 days / week, and on rare occasions my son's computer. All are connected by wifi. All do pretty mundane things: web browser, email, sometimes downloading patches, sometimes doing online backup. My 4 are running NTPD. 3 run windows most of the time and dual boot into linux. My 4th machine runs linux all the time. When my wife is here, she does a remote desktop type of thing into her work system. Two of my PC's are running Vista, one is running Windows 7. Those three dual boot into Ubuntu 11.04 and the always linux machine runs Ubuntu 11.04. Almost all the discussions I've had on this mailing list are for my Windows 7 machine. The path out of my house is: PC Wifi -- Wifi router -- wired router -- cable modem -- ISP -- internet - is this PC connected over wireless or wired? Wifi G - what else is going through the router? Someone else downloading large files or using streaming audio or video? See above. Normally, no huge data hogs. - who else might be sharing your connection? It's cable. Who knows? I also get cable TV and telephone through the same wire. - what type of service do you have? Presumably not dial-up! But what speed? Comcast Cable Just tested it with speedtest.net Ping to near city: 91 ms, Download: 29.63 Mbps, Upload: 5.3 Mbps - having checked you speed, would you describe your connection as stable? See above for speed. Generally, it's very stable. However, for the purposes we're discussing, I think it's latencies and delays that are the problem. As I mentioned in my reply to David L, I'm not concerned over trying to get stellar performance from internet servers. I just want to get a good GPS server system running and use the internet servers as a backup. Sincerely, Ron The quiet periods (e.g early Monday morning, Tuesday lunchtime) are much nearer to what I would hope for, which makes me wonder whether something is interfering with the connection outside those times. Cheers, David -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Hi David T, NOW you understand. I have 4 PC's connected to the LAN, plus my wife's work computer 3 days / week, and on rare occasions my son's computer. All are connected by wifi. All do pretty mundane things: web browser, email, sometimes downloading patches, sometimes doing online backup. My 4 are running NTPD. 3 run windows most of the time and dual boot into linux. My 4th machine runs linux all the time. When my wife is here, she does a remote desktop type of thing into her work system. Two of my PC's are running Vista, one is running Windows 7. Those three dual boot into Ubuntu 11.04 and the always linux machine runs Ubuntu 11.04. Almost all the discussions I've had on this mailing list are for my Windows 7 machine. The path out of my house is: PC Wifi -- Wifi router -- wired router -- cable modem -- ISP -- internet - is this PC connected over wireless or wired? Wifi G - what else is going through the router? Someone else downloading large files or using streaming audio or video? See above. Normally, no huge data hogs. - who else might be sharing your connection? It's cable. Who knows? I also get cable TV and telephone through the same wire. - what type of service do you have? Presumably not dial-up! But what speed? Comcast Cable Just tested it with speedtest.net Ping to near city: 91 ms, Download: 29.63 Mbps, Upload: 5.3 Mbps - having checked you speed, would you describe your connection as stable? See above for speed. Generally, it's very stable. However, for the purposes we're discussing, I think it's latencies and delays that are the problem. As I mentioned in my reply to David L, I'm not concerned over trying to get stellar performance from internet servers. I just want to get a good GPS server system running and use the internet servers as a backup. Sincerely, Ron Ron, Thanks for that clarification. I think you should be getting /very much/ better performance from your Internet servers. That ping is poor as well. Here I have 30 Mb/s down, just 1 Mb/s up, and NTP delays show as 18-34 ms (most in 22-30 ms). To that end, if your cable modem has multiple ports, connect one of the PCs direct to the CM and run it as your local NTP server. Wi-Fi doesn't help NTP. Later, you can add GPS/PPS to that PC as well. At the very least, connect your timekeeping PC direct to the wired router. No Wi-Fi! For my main NTP server, I got a low-powered and fan-less Intel Atom system, and it runs FreeBSD. It /only/ runs NTP, no interactive stuff at all. You might also consider getting rid of the two routers and just using the wireless one. You might also see whether you can run NTP on your router - perhaps it's a model which can run the DD-WRT firmware. I think some variants of DD-WRT can run NTP, but please check. I have a WRT 54GL in my system where I run the DD-WRT firmware. http://www.dd-wrt.com/site/index Online backup could well affect the performance of the connection for timekeeping, and your imminent Sure GPS/PPS will help enormously. Just some thoughts which may help you along the way. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. I have 4 PC's connected to the LAN, plus my wife's work computer 3 days / week, and on rare occasions my son's computer. All are connected by wifi. All do pretty mundane things: web browser, email, sometimes downloading patches, sometimes doing online backup. My 4 are running NTPD. 3 run windows most of the time and dual boot into linux. My 4th machine runs linux all the time. When my wife is here, she does a remote desktop type of thing into her work system. Two of my PC's are running Vista, one is running Windows 7. Those three dual boot into Ubuntu 11.04 and the always linux machine runs Ubuntu 11.04. Almost all the discussions I've had on this mailing list are for my Windows 7 machine. The path out of my house is: PC Wifi -- Wifi router -- wired router -- cable modem -- ISP -- internet - is this PC connected over wireless or wired? Wifi G - what else is going through the router? Someone else downloading large files or using streaming audio or video? See above. Normally, no huge data hogs. - who else might be sharing your connection? It's cable. Who knows? I also get cable TV and telephone through the same wire. - what type of service do you have? Presumably not dial-up! But what speed? Comcast Cable Just tested it with speedtest.net Ping to near city: 91 ms, Download: 29.63 Mbps, Upload: 5.3 Mbps - having checked you speed, would you describe your connection as stable? See above for speed. Generally, it's very stable. However, for the purposes we're discussing, I think it's latencies and delays that are the problem. As I mentioned in my reply to David L, I'm not concerned over trying to get stellar performance from internet servers. I just want to get a good GPS server system running and use the internet servers as a backup. Sincerely, Ron Ron, Thanks for that clarification. I think you should be getting /very much/ better performance from your Internet servers. That ping is poor as well. Here I have 30 Mb/s down, just 1 Mb/s up, and NTP delays show as 18-34 ms (most in 22-30 ms). To that end, if your cable modem has multiple ports, connect one of the PCs direct to the CM and run it as your local NTP server. Wi-Fi doesn't help NTP. Later, you can add GPS/PPS to that PC as well. At the very least, connect your timekeeping PC direct to the wired router. No Wi-Fi! For my main NTP server, I got a low-powered and fan-less Intel Atom system, and it runs FreeBSD. It /only/ runs NTP, no interactive stuff at all. You might also consider getting rid of the two routers and just using the wireless one. You might also see whether you can run NTP on your router - perhaps it's a model which can run the DD-WRT firmware. I think some variants of DD-WRT can run NTP, but please check. I have a WRT 54GL in my system where I run the DD-WRT firmware. http://www.dd-wrt.com/site/index Online backup could well affect the performance of the connection for timekeeping, and your imminent Sure GPS/PPS will help enormously. Just some thoughts which may help you along the way. Cheers, David Hi David, I really appreciate all these suggestions you shared, as well as past ones. If I decide to revamp my network, I'll probably put some of them into use. However, that's not really practical right now. All the networking gear is in the basement and all the PC's are upstairs. I've already spent 2 months working on this GPS stuff and ignoring some other things I need to be doing. I think I'm just going to focus on getting the Sure board up and running on a time server when it comes and, perhaps, use my BU-353 as a backup time source, and use the internet servers as a third level backup source. Most of the time, I won't care what they're doing Hopefully, I can spend less time thinking about time and just know that my PC's clocks are right. Regardless, I'm going to continue to have an interest in the topic and have enjoyed all the discussions and learning that have occurred. It's just that I have to do at least a few other things besides this. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/14/2012 5:04 PM, Ron Frazier (NTP) wrote: On 3/14/2012 4:00 PM, David J Taylor wrote: Hi David T, NOW you understand. I have 4 PC's connected to the LAN, plus my wife's work computer 3 days / week, and on rare occasions my son's computer. All are connected by wifi. All do pretty mundane things: web browser, email, sometimes downloading patches, sometimes doing online backup. My 4 are running NTPD. 3 run windows most of the time and dual boot into linux. My 4th machine runs linux all the time. When my wife is here, she does a remote desktop type of thing into her work system. Two of my PC's are running Vista, one is running Windows 7. Those three dual boot into Ubuntu 11.04 and the always linux machine runs Ubuntu 11.04. Almost all the discussions I've had on this mailing list are for my Windows 7 machine. The path out of my house is: PC Wifi -- Wifi router -- wired router -- cable modem -- ISP -- internet - is this PC connected over wireless or wired? Wifi G - what else is going through the router? Someone else downloading large files or using streaming audio or video? See above. Normally, no huge data hogs. - who else might be sharing your connection? It's cable. Who knows? I also get cable TV and telephone through the same wire. - what type of service do you have? Presumably not dial-up! But what speed? Comcast Cable Just tested it with speedtest.net Ping to near city: 91 ms, Download: 29.63 Mbps, Upload: 5.3 Mbps - having checked you speed, would you describe your connection as stable? See above for speed. Generally, it's very stable. However, for the purposes we're discussing, I think it's latencies and delays that are the problem. As I mentioned in my reply to David L, I'm not concerned over trying to get stellar performance from internet servers. I just want to get a good GPS server system running and use the internet servers as a backup. Sincerely, Ron Ron, Thanks for that clarification. I think you should be getting /very much/ better performance from your Internet servers. That ping is poor as well. Here I have 30 Mb/s down, just 1 Mb/s up, and NTP delays show as 18-34 ms (most in 22-30 ms). To that end, if your cable modem has multiple ports, connect one of the PCs direct to the CM and run it as your local NTP server. Wi-Fi doesn't help NTP. Later, you can add GPS/PPS to that PC as well. At the very least, connect your timekeeping PC direct to the wired router. No Wi-Fi! For my main NTP server, I got a low-powered and fan-less Intel Atom system, and it runs FreeBSD. It /only/ runs NTP, no interactive stuff at all. You might also consider getting rid of the two routers and just using the wireless one. You might also see whether you can run NTP on your router - perhaps it's a model which can run the DD-WRT firmware. I think some variants of DD-WRT can run NTP, but please check. I have a WRT 54GL in my system where I run the DD-WRT firmware. http://www.dd-wrt.com/site/index Online backup could well affect the performance of the connection for timekeeping, and your imminent Sure GPS/PPS will help enormously. Just some thoughts which may help you along the way. Cheers, David Hi David, I really appreciate all these suggestions you shared, as well as past ones. If I decide to revamp my network, I'll probably put some of them into use. However, that's not really practical right now. All the networking gear is in the basement and all the PC's are upstairs. I've already spent 2 months working on this GPS stuff and ignoring some other things I need to be doing. I think I'm just going to focus on getting the Sure board up and running on a time server when it comes and, perhaps, use my BU-353 as a backup time source, and use the internet servers as a third level backup source. Most of the time, I won't care what they're doing Hopefully, I can spend less time thinking about time and just know that my PC's clocks are right. Regardless, I'm going to continue to have an interest in the topic and have enjoyed all the discussions and learning that have occurred. It's just that I have to do at least a few other things besides this. Sincerely, Ron PS to my prior message. I don't think the problem so much is the delay to the internet servers, or even to get out of my house. NTPD is supposed to take care of that as long as it's pretty much symmetrical. I think the problem is that the Windows clock is like a wild tiger that doesn't want to be tamed and which is running every which way. For whatever reason, cpu load, heat, cosmic vibrations, whatever, the intrinsic frequency of the windows clock is always changing. In order to avoid beating up on the internet servers too much, I have to poll them at least every 4 minutes apart. If you let it, NTPD will extend that out to 16 minutes or more. So, when the clock source is polled, say the PC
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) wrote: My typical performance from internet servers is offsets of + / - 60 ms. # Why not try some that have a closer network distance, or are at least geographically closer? # Georgia / Comcast Hints: tos cohort 1 restrict source nomodify pool us.pool.ntp.org iburst # supposed to give you geographically close servers # server rolex.usg.edu iburst # server timex.usg.edu iburst # server ntp2.stsn.net iburst # server tick.gatech.edu iburst # server time.intersecur.net iburst # server nist1-atl.ustiming.org iburst # server nist1.columbiacountyga.gov iburst # _Your_ ISP router is likely already picking up the closest couple of these all by itself already # server ntp01.comcastbusiness.net iburst # server ntp02.comcastbusiness.net iburst # server ntp.whm.comcastcommercial.net iburst # server ntp01.inflow.pa.bo.comcast.net iburst # server ntp02.inflow.pa.bo.comcast.net iburst # server ntp01.cmc.co.denver.comcast.net iburst # server ntp02.cmc.co.denver.comcast.net iburst # server cr01.atlanta.ga.ibone.comcast.net iburst # server pr01.atlanta.ga.ibone.comcast.net iburst # server pe02.56marietta.ga.ibone.comcast.net iburst # server te-0-1-0-4-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-3-5-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-3-5-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-1-4-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-2-2-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-1-5-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-4-3-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-0-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-0-10-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-3-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-4-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-4-12-0-0-cr01.atlanta.ga.ibone.comcast.net iburst # server pos-0-0-0-0-pe01.56marietta.ga.ibone.comcast.net iburst # ... -- E-Mail Sent to this address blackl...@anitech-systems.com will be added to the BlackLists. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Any chance these PCs are using a timesource (in the system somewhere) that isn't constant in the face of power management events? The other pink tiger (vs elephant) in your house could/would be the wifi - at least in terms of being a stable network type. Some of the bufferbloat http://www.bufferbloat.net/ folks have little nice to say about wireless. rick jones http://www.netperf.org/ -- oxymoron n, Hummer H2 with California Save Our Coasts and Oceans plates these opinions are mine, all mine; HP might not want them anyway... :) feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH... ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. [] My system is Windows 7, NTP 4.2.7p259, GPS GlobalSat BU-353 USB NMEA only with GPGGA sentence at 57,600 baud. If anyone can shed some light on what happened, please do. It's driving me bonkers. I don't believe the GPS is at fault, and I suspect NTPD. Thanks in advance. Sincerely, Ron As all the Internet servers agreed (in another graph you showed me), I would suggest that you stick with the Internet servers and discard that GPS! I hope it will be better with the PPS from the Sure board. At least remove the noselect from the Internet servers so that NTP can do its job properly. At the moment you are telling NTP to trust the GPS no matter what it says. Can't think that sounds the wisest option. I might be inclined to add noselect to the GPS and measure it against the Internet servers. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On Tue, Mar 13, 2012 at 9:52 AM, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looks to me like that GPS, maybe lock lock and was able to coast on hold over for a while then fell off a cliff. If you have a log of satellite signal to noise ratios you might be able to figure out why. The internet servers appear to be 100% reliable seeing as they all agree. These kinds of things are why some hobbysts end up buying multiple GPS (different brands) Otherwise it is hard to sort out a GPS firmware bug from a solar storm or just that there were not sats visable to your indoor antenna for a few minutes I think your goal is to learn about all of this so these problems are a good thing. No one learns much from working systems. But if the goal is a reliable NTP server, the pool NTP servers can't be beat except by a good timing mode GPS, that has good self diagnostics and PPS. The self diagnostics part is important -- Chris Albertson Redondo Beach, California ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f5f9233.1080...@c3energy.com... [] The question is, what reset my clock abruptly 50 seconds off at around 0637 UTC. I would suspect your GPS. My loopstats file shows that the GPS was humming along quite nicely at that point with no anomalies at all. I don't believe it's the cause of the problem. It looks like NTPD abruptly decided to step my clock 50 seconds, for reasons unknown. As far as I know, NTPD is the only thing active on the system that will alter the clock. The Windows internet time function is disabled. The noselects are there for testing only. I won't keep them there permanently. I'm trying to determine where these anomalous readings are coming from. Sincerely, Ron Yes, I appreciate you are testing, but I think you are testing the wrong way round (in view of the results). Lock your PC as best you can to the Internet servers, and watch the GPS with noselect against it. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) wrote: I put the following in ntp.conf to allow the clock to coast in local mode if the GPS goes insane. # Allow system to coast on local clock if GPS is insane server 127.127.1.1 minpoll 3 maxpoll 6 # LCL, local clock fudge 127.127.1.1 stratum 12 # increase stratum I can't see how that will make any difference if the GPS is still selected, as your meinberg billboard seemed to indicate. http://dl.dropbox.com/u/9879631/drifting02a%20-%20insane.jpg ... and you would likely be better off using orphan than local clock, (not that it would make any diff if the GPS was still selected). -- E-Mail Sent to this address blackl...@anitech-systems.com will be added to the BlackLists. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 4:27 PM, David J Taylor wrote: Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote in message news:4f5f9233.1080...@c3energy.com... [] The question is, what reset my clock abruptly 50 seconds off at around 0637 UTC. I would suspect your GPS. My loopstats file shows that the GPS was humming along quite nicely at that point with no anomalies at all. I don't believe it's the cause of the problem. It looks like NTPD abruptly decided to step my clock 50 seconds, for reasons unknown. As far as I know, NTPD is the only thing active on the system that will alter the clock. The Windows internet time function is disabled. The noselects are there for testing only. I won't keep them there permanently. I'm trying to determine where these anomalous readings are coming from. Sincerely, Ron Yes, I appreciate you are testing, but I think you are testing the wrong way round (in view of the results). Lock your PC as best you can to the Internet servers, and watch the GPS with noselect against it. Cheers, David Hi David, OK, you've convinced me. I've put noselect on the GPS and taken it away on the New York NIST server. The NY NIST server is now preferred and polling from 1 - 4 minutes. Hopefully they won't ban me for hitting it too often. I think the NIST server will have more jitter, particularly as the polling interval increases, but we'll see what happens. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
Ron Frazier (NTP) wrote: On 3/13/2012 2:40 PM, Chris Albertson wrote: On Tue, Mar 13, 2012 at 9:52 AM, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looks to me like that GPS, maybe lock lock and was able to coast on hold over for a while then fell off a cliff. If you have a log of satellite signal to noise ratios you might be able to figure out why. The internet servers appear to be 100% reliable seeing as they all agree. These kinds of things are why some hobbysts end up buying multiple GPS (different brands) Otherwise it is hard to sort out a GPS firmware bug from a solar storm or just that there were not sats visable to your indoor antenna for a few minutes I think your goal is to learn about all of this so these problems are a good thing. No one learns much from working systems. But if the goal is a reliable NTP server, the pool NTP servers can't be beat except by a good timing mode GPS, that has good self diagnostics and PPS. The self diagnostics part is important The link you quoted just now is not the one that bugs me so much today. This one is the one that bugs me today, where my clock was stepped by 50 seconds for some reason. http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg I'm on a text mostly system so why do you require me to use a web browser to help you? What is your current ntp.conf, ntpd version, operating system and response of ntpq -crv -p after ntpd has had at least a day to stabilise? Not so long ago you posted that you had an offset within +/- 6 ms but didn't give any evidence. Response from ntpq -crv -p or better a day from peer_summary or couple of days from loop_summary would be enough. David When I get the Sure board, I plan to hang it on the pc in tandem with this GPS and compare. Sincerely, Ron ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 5:01 PM, E-Mail Sent to this address will be added to the BlackLists wrote: Ron Frazier (NTP) wrote: I put the following in ntp.conf to allow the clock to coast in local mode if the GPS goes insane. # Allow system to coast on local clock if GPS is insane server 127.127.1.1 minpoll 3 maxpoll 6 # LCL, local clock fudge 127.127.1.1 stratum 12 # increase stratum I can't see how that will make any difference if the GPS is still selected, as your meinberg billboard seemed to indicate. http://dl.dropbox.com/u/9879631/drifting02a%20-%20insane.jpg ... and you would likely be better off using orphan than local clock, (not that it would make any diff if the GPS was still selected). At that time, I had the GPS as the only selectable clock, for testing. I'm monitoring the internet servers for comparison. I've since selected 1 NIST server as the preferred and only selectable clock, and am monitoring the GPS and other internet clocks for comparison. My thinking was simply that if all the available clocks vanished (internet down), or the GPS went insane, I just want the system to coast on its internal clock until a viable clock source returns. I figure it may gain or lose up to 10 ms / minute this way, but at least it shouldn't do anything radical like jumping 50 seconds. I looked at the docs page for the orphan command for a few minutes and my eyes just glazed over. I decided to try the local clock option instead for now. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 5:39 PM, David Lord wrote: Ron Frazier (NTP) wrote: On 3/13/2012 2:40 PM, Chris Albertson wrote: On Tue, Mar 13, 2012 at 9:52 AM, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Looks to me like that GPS, maybe lock lock and was able to coast on hold over for a while then fell off a cliff. If you have a log of satellite signal to noise ratios you might be able to figure out why. The internet servers appear to be 100% reliable seeing as they all agree. These kinds of things are why some hobbysts end up buying multiple GPS (different brands) Otherwise it is hard to sort out a GPS firmware bug from a solar storm or just that there were not sats visable to your indoor antenna for a few minutes I think your goal is to learn about all of this so these problems are a good thing. No one learns much from working systems. But if the goal is a reliable NTP server, the pool NTP servers can't be beat except by a good timing mode GPS, that has good self diagnostics and PPS. The self diagnostics part is important The link you quoted just now is not the one that bugs me so much today. This one is the one that bugs me today, where my clock was stepped by 50 seconds for some reason. http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg I'm on a text mostly system so why do you require me to use a web browser to help you? What is your current ntp.conf, ntpd version, operating system and response of ntpq -crv -p after ntpd has had at least a day to stabilise? Not so long ago you posted that you had an offset within +/- 6 ms but didn't give any evidence. Response from ntpq -crv -p or better a day from peer_summary or couple of days from loop_summary would be enough. David When I get the Sure board, I plan to hang it on the pc in tandem with this GPS and compare. Sincerely, Ron Hi David L., I just figured the images and dropbox would be the fastest way to distribute the information. I didn't think I could attach files to messages going to the NTP questions mailing list. I'll be glad to provide stats files if anyone is interested. Just tell me how to distribute them. Based on advice from David Taylor, I just changed my configuration. Previously, I had only the GPS selectable and the internet servers noselected for testing purposes to try to determine where a slow drifting behavior is coming from. Now, I have changed that so the GPS is noselected and the New York NIST server is the preferred and only selectable peer. I am monitoring the GPS and other internet servers for comparison. So, it will be a couple of days before this configuration accumulates some stats files. Even though I have been tinkering with this GPS for a couple of months, I have never seen anything like the 50 second jump I started this thread about. That problem may not be reproducible for some time, if at all. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On Mar 13, 2012, at 3:19 PM, Ron Frazier (NTP) wrote: At that time, I had the GPS as the only selectable clock, for testing. I'm monitoring the internet servers for comparison. I've since selected 1 NIST server as the preferred and only selectable clock, and am monitoring the GPS and other internet clocks for comparison. ntpd is probably better able to judge which timeserver to use then forcing the choice via prefer, as it continues to evaluate the servers available to it and will use another one if the chosen peer happens to become unreachable. My thinking was simply that if all the available clocks vanished (internet down), or the GPS went insane, I just want the system to coast on its internal clock until a viable clock source returns. I figure it may gain or lose up to 10 ms / minute this way, but at least it shouldn't do anything radical like jumping 50 seconds. If you let ntpd run long enough to estimate the intrinsic drift, it will continue to adjust the clock even without external timesources. However, if your system clock really is off by a rate of 1:600 or worse, it's broken and ntpd probably wouldn't be able to fix it, at least without tinkering the max allowable slew rate-- running ntpdate via cron might do. I looked at the docs page for the orphan command for a few minutes and my eyes just glazed over. I decided to try the local clock option instead for now. Why would you want to use the local clock option? It's rarely the right solution, barring unusual circumstances... Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 6:42 PM, Chuck Swiger wrote: On Mar 13, 2012, at 3:19 PM, Ron Frazier (NTP) wrote: At that time, I had the GPS as the only selectable clock, for testing. I'm monitoring the internet servers for comparison. I've since selected 1 NIST server as the preferred and only selectable clock, and am monitoring the GPS and other internet clocks for comparison. ntpd is probably better able to judge which timeserver to use then forcing the choice via prefer, as it continues to evaluate the servers available to it and will use another one if the chosen peer happens to become unreachable. My thinking was simply that if all the available clocks vanished (internet down), or the GPS went insane, I just want the system to coast on its internal clock until a viable clock source returns. I figure it may gain or lose up to 10 ms / minute this way, but at least it shouldn't do anything radical like jumping 50 seconds. If you let ntpd run long enough to estimate the intrinsic drift, it will continue to adjust the clock even without external timesources. However, if your system clock really is off by a rate of 1:600 or worse, it's broken and ntpd probably wouldn't be able to fix it, at least without tinkering the max allowable slew rate-- running ntpdate via cron might do. I looked at the docs page for the orphan command for a few minutes and my eyes just glazed over. I decided to try the local clock option instead for now. Why would you want to use the local clock option? It's rarely the right solution, barring unusual circumstances... Regards, I was speculating that perhaps my only selectable clock, the GPS, failed, and that something went nuts, and that's why I found the clock 50 seconds off this morning. However, I don't have any evidence of a GPS failure. In any case, I just figured the local option might prevent any major clock changes if all other sources are not available. I only wanted the local option to kick in if there were no other sources. I did not have that option in the ntp.conf file when I started this thread. When I started this thread, my GPS was the only selectable clock, and there was no local option. For now, I'm mainly wanting to compare the GPS to one primary other source because I've been experiencing a slow drift in NMEA time with a variation of about 120 ms and an oscillation period of several days. I'm trying to isolate the source, either the GPS, or the subsystem that's getting time from the internet servers. I'm assuming all the internet servers are not drifting, but my instrumentation, ie the Meinberg time server monitor, or whatever drives it, could be off. That's why only have one source selectable, either the GPS, or one NIST server. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On Mar 13, 2012, at 4:08 PM, Ron Frazier (NTP) wrote: I was speculating that perhaps my only selectable clock, the GPS, failed, and that something went nuts, and that's why I found the clock 50 seconds off this morning. However, I don't have any evidence of a GPS failure. In any case, I just figured the local option might prevent any major clock changes if all other sources are not available. I only wanted the local option to kick in if there were no other sources. I did not have that option in the ntp.conf file when I started this thread. When I started this thread, my GPS was the only selectable clock, and there was no local option. One problem is that the local clock will always claim to be perfectly sync'ed, regardless of how close or far from real time it actually happens to be. The fact that ntpd's stats will look wonderful doesn't mean anything in this case. If you configured a reasonable # of internet NTP servers, and maybe your GPS in noselect mode while you try to debug whatever the issue is with it, you'd likely avoid a number of the issues you've been reporting. For now, I'm mainly wanting to compare the GPS to one primary other source because I've been experiencing a slow drift in NMEA time with a variation of about 120 ms and an oscillation period of several days. I'm trying to isolate the source, either the GPS, or the subsystem that's getting time from the internet servers. I'm assuming all the internet servers are not drifting, but my instrumentation, ie the Meinberg time server monitor, or whatever drives it, could be off. That's why only have one source selectable, either the GPS, or one NIST server. Well, it's hardly unexpected that the NMEA sentences would show variable timing; they are only supposed to be correct to the nearest second, so 120 ms is well within specs. It's the PPS signal which is expected to provide sub-millisecond precision. [ Hasn't this already been discussed? ] Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 7:24 PM, Chuck Swiger wrote: On Mar 13, 2012, at 4:08 PM, Ron Frazier (NTP) wrote: I was speculating that perhaps my only selectable clock, the GPS, failed, and that something went nuts, and that's why I found the clock 50 seconds off this morning. However, I don't have any evidence of a GPS failure. In any case, I just figured the local option might prevent any major clock changes if all other sources are not available. I only wanted the local option to kick in if there were no other sources. I did not have that option in the ntp.conf file when I started this thread. When I started this thread, my GPS was the only selectable clock, and there was no local option. One problem is that the local clock will always claim to be perfectly sync'ed, regardless of how close or far from real time it actually happens to be. The fact that ntpd's stats will look wonderful doesn't mean anything in this case. If you configured a reasonable # of internet NTP servers, and maybe your GPS in noselect mode while you try to debug whatever the issue is with it, you'd likely avoid a number of the issues you've been reporting. For now, I'm mainly wanting to compare the GPS to one primary other source because I've been experiencing a slow drift in NMEA time with a variation of about 120 ms and an oscillation period of several days. I'm trying to isolate the source, either the GPS, or the subsystem that's getting time from the internet servers. I'm assuming all the internet servers are not drifting, but my instrumentation, ie the Meinberg time server monitor, or whatever drives it, could be off. That's why only have one source selectable, either the GPS, or one NIST server. Well, it's hardly unexpected that the NMEA sentences would show variable timing; they are only supposed to be correct to the nearest second, so 120 ms is well within specs. It's the PPS signal which is expected to provide sub-millisecond precision. [ Hasn't this already been discussed? ] Regards, I don't have PPS on this computer. It has no serial port. When I get my Sure board, assuming I can solder it without killing it, I'm going to try bringing the PPS signal through a Prolific chipset based serial - USB converter which is supposed to pass handshaking signals. I don't know if that will work. Of course, this thread was started because of a mysterious 50 second jump in my system clock. That's still a mystery. Regarding the NMEA though. If I can get those data sentences to my PC with a jitter of 3 ms through USB, which I can, I don't see why the start and end time of the sentence should vary by 120 ms over 2-4 days. I'm trying to figure out where that variance is coming from. If I can consistently keep time with this device, or some other USB device, to within + / - 6 ms or even 10 ms of true time, that's plenty good enough for my purposes. If I can only keep time to within + / - 70 ms, then the device is more or less worthless, no better than the internet servers. So, I'm trying to figure out whether it's good enough, or worthless. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On Mar 13, 2012, at 5:05 PM, Ron Frazier (NTP) wrote: I don't have PPS on this computer. It has no serial port. When I get my Sure board, assuming I can solder it without killing it, I'm going to try bringing the PPS signal through a Prolific chipset based serial - USB converter which is supposed to pass handshaking signals. I don't know if that will work. You should be able to get reasonable time via USB, but not of the quality that a serial or parallel connection would provide, especially one where the ISR does timestamping. Of course, this thread was started because of a mysterious 50 second jump in my system clock. That's still a mystery. Yes. I've seen that sort of thing with VMs; some older Linux kernels also had some fairly bad timekeeping issues. Regarding the NMEA though. If I can get those data sentences to my PC with a jitter of 3 ms through USB, which I can, I don't see why the start and end time of the sentence should vary by 120 ms over 2-4 days. I'm trying to figure out where that variance is coming from. If I can consistently keep time with this device, or some other USB device, to within + / - 6 ms or even 10 ms of true time, that's plenty good enough for my purposes. If I can only keep time to within + / - 70 ms, then the device is more or less worthless, no better than the internet servers. So, I'm trying to figure out whether it's good enough, or worthless. You shouldn't have too much difficulty getting timekeeping accuracy to within ~6 ms using public internet NTP servers which are nearby, although if you just use the global ntp pool and get some servers which are not very close, that might be on the order of +/- ~20 ms instead. Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 2012-03-13, Ron Frazier (NTP) timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. Did you shut down and restart your computer? Did you perchance do this during the daylight savings time transition on a Windows system? Could the error be related to the fact that Windows like time on localtime not UTC? http://dl.dropbox.com/u/9879631/drifting01-peerstats.20120312.jpg Here is what was showing on the Meinberg Time Server Monitor when I woke up: http://dl.dropbox.com/u/9879631/drifting02a%20-%20insane.jpg And the graph of the peerstats for that time: http://dl.dropbox.com/u/9879631/drifting02b%20-%20peerstats%20insane.jpg The clock error was REAL, as confirmed by my atomic wrist watch. However, the loopstats graph for the same time period shows no problem with the GPS: http://dl.dropbox.com/u/9879631/drifting02c%20-%20gps%20NOT%20insane.jpg So, I shut down NTPD and reset the time with a batch file that calls ntpdate and querys the New York NIST server: http://dl.dropbox.com/u/9879631/drifting02d%20-%20set%20with%20ntpdate.jpg Here is the time server monitor shortly after NTPD restart: http://dl.dropbox.com/u/9879631/drifting02e%20-%20shortly%20after%20ntpd%20restart.jpg And after a 2nd restart: http://dl.dropbox.com/u/9879631/drifting02f%20-%202nd%20ntpd%20restart%20after%20insane.jpg And here are the current peerstats, which look normal. The offset to the internet servers tends to drift and will eventually cross the zero line and get positive. http://dl.dropbox.com/u/9879631/drifting02g%20-%20peerstats%20after%20insane%20and%202nd%20reset.jpg The GPS appears to have been stable all through this, and was never powered off or unplugged. It looks like NTPD went crazy and reset my clock for some reason. Here are the peerstats and loopstats during the insane period. http://dl.dropbox.com/u/9879631/loopstats.20120313-1-restart%20around%201350%20utc http://dl.dropbox.com/u/9879631/peerstats.20120313-1-insane Here is my current ntp.conf: http://dl.dropbox.com/u/9879631/ntp.conf My system is Windows 7, NTP 4.2.7p259, GPS GlobalSat BU-353 USB NMEA only with GPGGA sentence at 57,600 baud. If anyone can shed some light on what happened, please do. It's driving me bonkers. I don't believe the GPS is at fault, and I suspect NTPD. Thanks in advance. Sincerely, Ron ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 10:05 PM, unruh wrote: On 2012-03-13, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. Did you shut down and restart your computer? Did you perchance do this during the daylight savings time transition on a Windows system? Could the error be related to the fact that Windows like time on localtime not UTC? Good question, but no. The change to DST occurred early Sunday morning. This event occurred early Monday morning and I noticed it today, Tuesday morning. By that time, the system was already well stabilized on DST. Sincerely, Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] ARRGH!!! I woke up to a 50 SECOND clock error.
On 3/13/2012 10:28 PM, Ron Frazier (NTP) wrote: On 3/13/2012 10:05 PM, unruh wrote: On 2012-03-13, Ron Frazier (NTP)timekeepingntpl...@c3energy.com wrote: Hi all, I just woke up to a 50 SECOND clock error. Prior to the error, with my PC locked into the GPS and the internet servers noselected, here's what my peerstats looked like. Baseline is the GPS. Colored lines are internet servers. Did you shut down and restart your computer? Did you perchance do this during the daylight savings time transition on a Windows system? Could the error be related to the fact that Windows like time on localtime not UTC? Good question, but no. The change to DST occurred early Sunday morning. This event occurred early Monday morning and I noticed it today, Tuesday morning. By that time, the system was already well stabilized on DST. Sincerely, Ron Scratch that last comment. I said it wrong. The 50 second event occurred early today, Tuesday morning, and I noticed it later today, Tuesday morning. Ron -- (PS - If you email me and don't get a quick response, don't be concerned. I get about 300 emails per day from alternate energy mailing lists and such. I don't always see new messages very quickly. If you need a reply and have not heard from me in 1 - 2 weeks, send your message again.) Ron Frazier timekeepingdude AT c3energy.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions