Re: [ntp:questions] Using NTP to calibrate sound app
In article 51033f49.215309...@news.eternal-september.org, no-...@no-place.org wrote: I am an app developer who has a precision audio frequency app for iPhone and Android devices. For my app the nominal crystal oscillator accuracy in these devices is not sufficient. Up until now I have been providing frequency calibration in my app by instructing the user to call the telephone feed of WWV (NIST) audio (using a separate landline phone) and let my app listen to the 500 Hz and 600 Hz tones. By analyzing the audio I can correct for the device's audio system clock deviation. Normally they only need to do this once after the app is installed because the stability of these devices is OK once I memorize the offset. Now I am considering an alternate means of performing this calibration using NTP. The iPhone and Android devices deliver audio to my app in small packets. A calibration run would consist of an initial NTP syncronization with an audio packet, followed by a period of some number of minutes during which I will just count audio packets, followed by a final NTP synchronization with the last audio packet. By knowing the time difference over some number of audio packets I hope to calculate the actual audio clock frequency for that device. My question is about the NTP procedure I should follow to do this. I obviously don't want to hard-code for a specific time server because things could change after the user gets my app and it is unfair to send a whole block of users to the same server. The Server Pool looks promising. Does pool.ntp.org just behave like a Stratum 2 server so I could hard-code that URL into my implementation of NTP in my app? I would appreciate any observations on the promise of this approach. How accurately do you need to calibrate? What will people use this app for? And how long can this calibration take before users revolt? Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTP vs RADclock?
In article jqgakc$420$1...@dont-email.me, David Woolley david@ex.djwhome.demon.invalid wrote: skillz...@gmail.com wrote: Has there been any independent comparison of NTP vs RADclock [1]? Information on the RADclock site seem to indicate it performs pretty well, but I haven't seen any analysis except from the RADclock authors. This is the first I've heard of it, so I assume that it has never appeared on this newsgroup, and its authors are not active here. What really annoys me, though, is that it fails to describe the essence of the algorithm on the first page. That's par for the course for commercial software, but this says that it is open source. Can you point me to where this information is provided (I think ntpd has a similar problem, though). I'm using NTP today for synchronization between devices on a LAN to each other of an internal clock separate from the system's normal wall clock (which uses the system's NTP). So I have flexibility with update intervals (1-3 seconds acceptable in my case) and other parameters. [1] http://www.synclab.org/radclock/ I looked at one of their demonstrations, TIM_2008_camera.pdg, where they showed 10 microsecond sync over a LAN, with very bad NTP performance. But I've done 7 microsconds with NTP in a quiet testbed on 1996-era hardware, so I don't know what the problem with NTP was. I'll have to read the paper more carefully. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] local refclock and orphan mode...
In article e1rt3rk-0005bs...@stenn.ntp.org, Harlan Stenn st...@ntp.org wrote: Doug, On 11/21/2011 01:51 AM, Harlan Stenn wrote: I asked this on hackers@ and think a wider audience would be good. In the old days, we had the local refclock. Now we have orphan mode. Can anybody think of a (good) use-case where one would want *both* the local refclock *and* orphan mode configured for an instance of ntpd? Harlan, I think you are confused about how things operate around here. The way this works is that *we* ask you the questions and then *you* give us answers;) Mostly, that's true :) Mostly... In this case I'm trying to make sure we don't implement a change that would catch folks by surprise. The first thing that comes to mind is environments with mixed ntpd versions. I realize that pre-4.2.2 was 5 years ago but sometimes change management policies read more like change resistance policies. Sure, but in the old days there was no orphan mode. And there have been some other changes to things that would require updating ntp.conf files. So while you mention good points, I'm still not hearing anybody say We use local refclocks *and* orphan mode and the reason for that is X and here's how we expect it to behave. What about the infamous interstellar/interplanetary ntp network? If they are up for upgrading their ntpd instances, they can easily upgrade the config files at the same time. If such networks have what they think is a valid use case for simultaneous use of both local refclocks and orphan mode, it would be Good to hear what that case is. I may not be understanding the question correctly, but one similar-sounding real-world application comes to mind: There is a planned radar system where a GPS receiver distributes GPS System Time via IRIG-B. There are a number of Intel x86-64 servers running RHEL (with MRG) supporting realtime applications software. It is necessary that this applications code be synchronized to GPS System Time, so applications can stay in synch with the radar hardware command pipeline. One way to achieve this is to provide an IRIG receiver card, the Linux I/O driver needed to access the IRIG card, and a compatible reference clock driver compiled into the NTPv4 daemon. (Symmetricom offers such a triplet; there may be others as well.) This allows application code to use ordinary kernel-provided timers to trigger software to make the donuts exactly when needed. An alternative approach, successfully used on prior radars, is to have the IRIG card (or custom equivalent hardware) generate hardware timing interrupts, which interrupts are turned into UNIX signals sent to the application code. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] local refclock and orphan mode...
In article e1rtlis-0007ax...@stenn.ntp.org, Harlan Stenn st...@ntp.org wrote: Joe wrote: Harlan wrote: If such networks have what they think is a valid use case for simultaneous use of both local refclocks and orphan mode, it would be Good to hear what that case is. I may not be understanding the question correctly, but one similar-sounding real-world application comes to mind: There is a planned radar system where a GPS receiver distributes GPS System Time via IRIG-B. There are a number of Intel x86-64 servers running RHEL (with MRG) supporting realtime applications software. It is necessary that this applications code be synchronized to GPS System Time, so applications can stay in synch with the radar hardware command pipeline. One way to achieve this is to provide an IRIG receiver card, the Linux I/O driver needed to access the IRIG card, and a compatible reference clock driver compiled into the NTPv4 daemon. (Symmetricom offers such a triplet; there may be others as well.) This allows application code to use ordinary kernel-provided timers to trigger software to make the donuts exactly when needed. An alternative approach, successfully used on prior radars, is to have the IRIG card (or custom equivalent hardware) generate hardware timing interrupts, which interrupts are turned into UNIX signals sent to the application code. If there are multiple machines with these refclocks then just mesh them together, and in this case I don't see why either orphan mode or the local refclock is needed. There are two machines with IRIG connections and refclock drivers et al, and a factor more servers with only NTP exchange via ethernet. By mesh them together, what do you mean? If I'm missing something, I can see that in the old days a local refclock would have been good to keep the machines sync'd together, and now orphan mode would do the same thing. I do not see that in this case that *both* the local refclock driver *and* orphan mode would be useful. That was my suspicion, but of course this entire network is isolated from the outside world, with only GPS to keep time by. Maybe I'm unclear on the precise definition of orphan mode?. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] How to keep fake time in past/future?
In article slrnirobu0.8be.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2011-04-30, Joseph Gwinn joegw...@comcast.net wrote: In article slrnirllcv.g22.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2011-04-29, Cristian Seres cristia...@contrasec.fi wrote: Hi! How would you implement an NTP server which would need to offer a time set deliberately in past/future, say 365*86400 seconds, or even better - first set the freely chosen date on NTP server and then keep the hours, minutes and seconds in sync with the real time? Perhaps you could tell us why in the world you would want to do that? It's very common, actually. For instance, in Air Traffic Control they record everything, and later play prior events back, to figure out what happened. Alternately, very large and complex training scenarios are written, the scenario happening at some time different from the present, perhaps past, perhaps future. And why do we not hear from Seres who could tell us why, instead of us all guessing here. Any of the guesses I have seen have been trivially satisfied by simply killing ntpd and resetting the computer's clock. The OP wanted the milliseconds to be right, but the days or years wrong. Or if you want set up one computer with LOCAL clock as only source, and set the the inappropriate time, (without any other source of time) reset the clock and have the others use it as their server. But of course the milliseconds will not be right. But why are we not hearing from the person who presumably knows why they want their computer clock mistreated thus? Well, I don't know the details of that system, but for instance if one wished to test tolerance for leap seconds, it's necessary to abuse those hapless clocks. I've done this by manually resetting the clocks. This can really upset radar trackers if they aren't designed to tolerate the occasional one-second step discontinuity. Negative leaps are worse than positive leaps. As for simulation, probably the most common approach is to leave GPS, NTP, and the local computer clocks alone, instead interposing a software layer that can inject the needed constant offset. But this only works if one is able to install such a layer. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] How to keep fake time in past/future?
In article slrnirllcv.g22.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2011-04-29, Cristian Seres cristia...@contrasec.fi wrote: Hi! How would you implement an NTP server which would need to offer a time set deliberately in past/future, say 365*86400 seconds, or even better - first set the freely chosen date on NTP server and then keep the hours, minutes and seconds in sync with the real time? Perhaps you could tell us why in the world you would want to do that? It's very common, actually. For instance, in Air Traffic Control they record everything, and later play prior events back, to figure out what happened. Alternately, very large and complex training scenarios are written, the scenario happening at some time different from the present, perhaps past, perhaps future. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] UK - GPS Jamming update
In article AANLkTi=jmmuyufzkdj6pdfhgpadyhkeqkq-vqycjq...@mail.gmail.com, Dave Hart daveh...@gmail.com wrote: On Thu, Mar 31, 2011 at 3:27 PM, unruh un...@wormhole.physics.ubc.ca wrote: I think this is a really bad idea. It conveys the impression that it is OK to jam GPS-- Even the government thinks it is OK to jam GPS. This removes the (admittedly possibly small) moral argument that it is bad to jam GPS because of the harm it could do. This says Taking GPS off the air for 18 hours is fine. If you're within a few miles of one of the 40,000 planned LightSquared base stations, you'll likely lose GPS for good, if the FCC sticks to its current approval to repurpose a GPS-adjacent band allocated for space-based use for terrestrial broadband. Search for LightSquared or see for example: http://freegeographytools.com/2011/update-on-lightsquareds-gps-jamming-proposa l These LightSquared guys apparently know their way around the DC beltway, managing to schedule an accelerated approval to coincide with Thanksgiving. I can only suppose they bribed the right people. Did I say bribed? I'm sorry, expressed their first amendment rights via unlimited political donations! Silly me. The DoD has begun to push the FCC to protect GPS: Pentagon Raises Concerns About Lightsquared Wireless, Wall Street Journal, 31 March 2011. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] new driver development
In article j_mdnrcpita18hjqnz2dnuvz_tudn...@megapath.net, hal-use...@ip-64-139-1-69.sjc.megapath.net (Hal Murray) wrote: In article 2wlgp.34776$d46.31...@newsfe07.iad, Bruce Lilly bruce.li...@gmail.com writes: o POSIX mutex for synchronized access to shared memory for updates -- obviates mode 0 / mode 1 / OLDWAY I'm far from a POSIX wizard. When I google for POSIX mutex I get a bunch of hits that all are part of pthreads. Pthreads moved into POSIX, so no surprise. Does that stuff work across processes rather than threads? The mutex needs to be in shared memory so both processes can get at it. Right? Who initializes it? It would be best to read the actual standard. Mutexes spanning shared memory are supported, for exactly the reasons you list. POSIX.1: http://pubs.opengroup.org/onlinepubs/9699919799/. And the general history: http://en.wikipedia.org/wiki/POSIX Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] What level of timesynch error is typical on Win XP?
In article i9r6cb$62...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: Joseph Gwinn joegw...@comcast.net wrote in message news:joegwinn-ee48fd.22434621102...@news.giganews.com... In article i9pkvb$dc...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: [] You might consider providing a local, more precise NTP server with something like a small, fan-less Intel Atom system running FreeBSD and synched across the network to your GPS time server. You might be able to keep a small box like that in a more temperature controlled environment, but even without it might provide a way of smoothing out any jitter due to your remote connection to the GPS server. I'm not convinced that this would help. NTP reports a round trip time of slightly more than 2 mS, which is very close to the two milliseconds that ping sees, so it seems unlikely that the time server or intervening network is the root cause. [] Joe Gwinn No, I wasn't convinced either - hence it was just a suggestion. On the systems here, though, the NTP delay shows around 0.25-0.75 msec to the LAN servers. I must say that I don't know why ping sees 2 milliseconds, which did seem high, but I also don't know the physical location of the timeserver, or how many hops (and firewalls) it takes to get there. I'll have to explore it with traceroute. But the 2 ms RT time explains only a millisecond or so of timesync error, leaving much error to be explained. Research continues. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] What level of timesynch error is typical on Win XP?
In article i9njj8$ea...@news.eternal-september.org, David Woolley da...@ex.djwhome.demon.invalid wrote: Joseph Gwinn wrote: I have a small network of Windows XP (64 bit) running simulations, with NTPv4 running on all the boxes and using a GPS-based timeserver on the company network. The ping time to the server is 2 milliseconds from my desk, but I'm seeing random time errors of order plus/minus 5 to 10 milliseconds, based on loopstats data. Loopstats data cannot give you an accurate measure of error when ntpd is locked on. It will give a value that is distinctly pessimistic. If ntpd could measure the actual error, it ought to be possible for it to remove that error. This is certainly true, but loopstats and peerstats data is nonetheless useful as a proxy for the unmeasured actual clock offset. In other words, if loopstats data shows adequate stability for my application, true offset error will also be adequate. Unless the transport delay is asymmetric, which is not the case here. Nor do I have the hardware to make better offset measurements than those provided by loopstats and peerstats. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] What level of timesynch error is typical on Win XP?
In article i9oo4r$5a...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: David Woolley da...@ex.djwhome.demon.invalid wrote in message news:i9omts$7s...@news.eternal-september.org... Richard B. Gilbert wrote: Also note that Windows' clock ticks every 17 milliseconds. Only when not running ntpd. ntpd forces the use of multimedia timers. .. and not when running Windows-7 and possibly Vista, when it's just under 1 millisecond, and NTP uses the native timers. The platforms in question are running Windows XP, not Vista or Windows 7. How does this change the answer? By the way, the hardware is a collection of 8-core HP Z800 workstations connected together by copper gigabit ethernet links and local hub, with one link going via the company network to the GPS network time server. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] What level of timesynch error is typical on Win XP?
In article i9ooo9$9i...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: Joseph Gwinn joegw...@comcast.net wrote in message news:joegwinn-da4b7b.23340420102...@news.giganews.com... In article i9mqek$tr...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: I have a small network of Windows XP (64 bit) running simulations, with NTPv4 running on all the boxes and using a GPS-based timeserver on the company network. The ping time to the server is 2 milliseconds from my desk, but I'm seeing random time errors of order plus/minus 5 to 10 milliseconds, based on loopstats data. This level of timesynch error is OK for the simulation, but still that's a lot of error. I get far better on big UNIX boxes. The question is if this level of error is reasonable, given the setup. I know that timekeeping under Windows is not optimum, but cannot change the OS, so the question is if I have gotten things as good as they can be, or should I dig deeper. One thing that comes to mind is to raise the priority of the NTP daemon to exceed that of the simulation software. Thanks in advance, Joe Gwinn Joe, This is the performance I see: http://www.satsignal.eu/mrtg/performance_ntp.php The XP systems are: Feenix: GPS-synched Narvik: LAN-synced to Pixie (FreeBSD with GPS source) These are all over the place. Both hardware and OS seem to matter, by a lot. Hardly all over the place! Feenix is well within a milliseconds, and Narvik just within a millisecond, and programs on that OS can only read the system time with ~16ms precision. By all over the place I mean that while some combinations are very good, yielding peak offsets well less than a millisecond, some combinations yield peak offsets of 25 milliseconds. In my application, only peak offsets matter. I can't add a GPS source, and I can't really control temperature. So you need to keep the polling interval short. We tried 16 seconds, with no variation allowed, and it didn't make much difference. Currently, NTP is being allowed to choose its own polling period. I don't recall what periods it chose, but I'll look. What other periods would you suggest, and why? I don't think that iburst is the issue, because the randomness persists for at least a week, long after the iburst transients will have died down. I never said iburst was an issue, just that the systems will need to be on for several hours before best accuracy is achieved. It's a pity that NTP doesn't have a faster initial convergence. My experience is the same. for average behaviour. But for use in realtime, running the daemon at high realtime priority greatly reduces the tails of the probability distribution of response times and/or clock offsets. Joe Gwinn Yes, if the CPU loading is heavy I can quite believe that. That's the usual cause. My usual solution is to ensure that the NTP daemon has a high realtime priority that well exceeds that of the realtime application code. NTP can be run at the highest realtime priority available without difficulty on every system I have tried this on. Another, more subtle cause, is Network File System (NFS) access being used to read or write the local disk from afar. This completely distracts the local OS kernel, at an implied priority that exceeds all processes and threads, including NTP running at the highest realtime priority. And yet there may be no record of the activity in syslog. Nor is it clear that NFS activity is always counted in the I/O read and write statistics kept by the kernel. Diagnosis may require network tools, unless one can figure out where the NFS access must be coming from and stop it at the source. I should explain what I mean by the term realtime priority. There are two related but independent things going on here, a numerical priority and a scheduling policy. A realtime scheduling policy is typically winner take all, where the process (and/or thread) having the highest priority can use as much of the processor as desired, even if all other processes and threads are squeezed out completely. In other words, realtime scheduling policies are completely unfair. It is the human system designers' responsibility to ensure that there is enough computer that nothing critical is unduly stalled. A non-realtime scheduling policy attempts some notion of fairness, where all processes and threads make progress at an average rate that is determined by their respective numerical priorities. In such a scheme, nobody is completely squeezed out, and no direct human intervention is required to ensure this outcome. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] What level of timesynch error is typical on Win XP?
In article i9pkif$at...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: Evandro Menezes evan...@mailinator.com wrote in message news:a376dc23-cb31-441c-9b35-b10a9758c...@a36g2000yqc.googlegroups.com... [] Indeed, since Windows allows a process to be starved from running, depending on the load, a higher priority process may block NTP from running. Therefore, although raising the priority for NTP doesn't mean that it cannot be starved, it does decrease the likelihood of that happening. Linux, on the other hand, favors fair process scheduling and strives to not starve any from running at least for a little while. HTH Windows can run NTP at real-time priority, if you give the NTP user that right. Normal user processes will not then pre-empt NTP. We are trying this, but given that the error level and pattern showed no diurnal variation, I don't really expect changing priority to help. The simulations are run only during the day, so one would expect a diurnal variation if CPU load were the issue. We are also verifying the the Intel power-saving feature, which slows the CPU clock et al, is disabled. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
[ntp:questions] What level of timesynch error is typical on Win XP?
I have a small network of Windows XP (64 bit) running simulations, with NTPv4 running on all the boxes and using a GPS-based timeserver on the company network. The ping time to the server is 2 milliseconds from my desk, but I'm seeing random time errors of order plus/minus 5 to 10 milliseconds, based on loopstats data. This level of timesynch error is OK for the simulation, but still that's a lot of error. I get far better on big UNIX boxes. The question is if this level of error is reasonable, given the setup. I know that timekeeping under Windows is not optimum, but cannot change the OS, so the question is if I have gotten things as good as they can be, or should I dig deeper. One thing that comes to mind is to raise the priority of the NTP daemon to exceed that of the simulation software. Thanks in advance, Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] What level of timesynch error is typical on Win XP?
In article i9mqek$tr...@news.eternal-september.org, David J Taylor david-tay...@blueyonder.co.uk.invalid wrote: I have a small network of Windows XP (64 bit) running simulations, with NTPv4 running on all the boxes and using a GPS-based timeserver on the company network. The ping time to the server is 2 milliseconds from my desk, but I'm seeing random time errors of order plus/minus 5 to 10 milliseconds, based on loopstats data. This level of timesynch error is OK for the simulation, but still that's a lot of error. I get far better on big UNIX boxes. The question is if this level of error is reasonable, given the setup. I know that timekeeping under Windows is not optimum, but cannot change the OS, so the question is if I have gotten things as good as they can be, or should I dig deeper. One thing that comes to mind is to raise the priority of the NTP daemon to exceed that of the simulation software. Thanks in advance, Joe Gwinn Joe, This is the performance I see: http://www.satsignal.eu/mrtg/performance_ntp.php The XP systems are: Feenix: GPS-synched Narvik: LAN-synced to Pixie (FreeBSD with GPS source) These are all over the place. Both hardware and OS seem to matter, by a lot. Your best bet would be to add a GPS source to your Windows PC, when you might expect errors of less than 250 microseconds under stable running (i.e. leave the PC on 24 x 7). If you can't do that, PC Narvik suggests you might get within +/- 1.5ms. That's with a configuration file like: server A iburst maxpoll 5 server B iburst maxpoll 5 server C iburst maxpoll 5 where A, B and C all have a GPS source. All PCs on the same switch, so a much better ping than 2ms. You could reduce the maxpoll further to 4 (if the server operator agrees) and get somewhat better performance, and keeping the PCs in a stable temperature environment would also be likely to help. The bumps at 05:00 are when the heating comes on. I can't add a GPS source, and I can't really control temperature. I don't think that iburst is the issue, because the randomness persists for at least a week, long after the iburst transients will have died down. In my experience, changing the priority of NTP doesn't help a lot, but most of my PCs are not CPU-bound. But I have given the account the rights to do that. My experience is the same. for average behaviour. But for use in realtime, running the daemon at high realtime priority greatly reduces the tails of the probability distribution of response times and/or clock offsets. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Allan deviation survey
In article 1f42m7-7gr2@ntp.tmsw.no, Terje Mathisen terje.mathisen at tmsw.no wrote: Joseph Gwinn wrote: The address did not look munged to me either. It makes perfect sense for a physicist to name servers after physics objects. The standard approach is to put the demunging instructions in your sig. Like: Please remove reference to the entrance to a worm's burrow from email address. No, no! Please remove interstellar gateway from my address ? Something too hard for a computer to figure out, but easy for a human. :-) I like it. But will unruh? Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Allan deviation survey
In article slrni8sddh.n3i.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2010-09-13, Joseph Gwinn joegw...@comcast.net wrote: Unruh, In article slrni8ru62.i6p.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2010-09-13, David L. Mills mi...@udel.edu wrote: [snip] ... And, by the way, mail sent to your alleged mail address is returned to sender as undeliverable. Yes, I am sorry about that but it is done in order to slightly reduce the spam I get. It should be clear how to alter it, but I realise that that makes more work for the responder. For a long time I did not munge my address, and as a result am on a number of spam lists. The address did not look munged to me either. It makes perfect sense for a physicist to name servers after physics objects. Ah, I finally looked at it. I used to use the nn new reader which munged my email address. I recently (well a year ago) switched to slrn, and just assumed that the same would occur there. Your comments caused me to actually read one of my posts as it appeared on the newsgoup, and sure enough it is the address of the machine running slrn ( which does not receive mail) instead of the munged address. Sorry about the wrong explanation. If you really want to email me you can remove the wormhole. But answering on the list is probably better anyway. Ah. I wasn't today trying to email anyone, but I could see the problem should I try. Anyway, now that it's understood, it can be fixed. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Allan deviation survey
Unruh, In article slrni8ru62.i6p.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2010-09-13, David L. Mills mi...@udel.edu wrote: [snip] ... And, by the way, mail sent to your alleged mail address is returned to sender as undeliverable. Yes, I am sorry about that but it is done in order to slightly reduce the spam I get. It should be clear how to alter it, but I realise that that makes more work for the responder. For a long time I did not munge my address, and as a result am on a number of spam lists. The address did not look munged to me either. It makes perfect sense for a physicist to name servers after physics objects. The standard approach is to put the demunging instructions in your sig. Like: Please remove reference to the entrance to a worm's burrow from email address. Something too hard for a computer to figure out, but easy for a human. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] DATUM TymServe 2000 Op/Sv Manual request
In article eov7o.197415$9f6.393...@twister1.libero.it, mauri maremovebeforereplyuri...@libero.it wrote: In Symmetricom.com this model isn't available. Call Symmetricom up and ask them. Joe Gwinn E-Mail Sent to this address will be added to the BlackLists n...@blacklist.anitech-systems.invalid ha scritto nel messaggio news:i3hvbr$f2...@news.eternal-september.org... mauri wrote: I`m looking for DATUM TymServe 2000 (TS2000-GPS) Network Time Server Operating/Service manual - symmetricom.com ? Symmetricom and Datum merged in 2002. -- 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] IA approved COTS NTP servers question
In article 62b84ad9-7d4c-4074-960e-aae4ef826...@u7g2000yqm.googlegroups.com, Fran fran.ho...@jhuapl.edu wrote: On Jun 4, 3:13 pm, Greg Hennessy greg.henne...@cox.net wrote: On 2010-06-04, Fran fran.ho...@jhuapl.edu wrote: On Jun 3, 4:49?pm, Greg Hennessy greg.henne...@cox.net wrote: Do you know of any DISA IA approved COTS NTP servers ? Why not use tick.usno.navy.mil or tock.usno.navy.mil? Only half a smiley. Thats a funny one Greg, thanks! On the serious side, if you are worried about having to follow DISA STIGS, then it seems safe to assume you are on NIPR or SIPR nets, in which case it is probably easier to use the USNO supplied time service rather than recreating your own. If for redundancy you wish to run your own NTP servers (which you should point to USNO since USNO is what all DoD sources are *SUPPOSED* to be using), I'm not aware of any COTS NTP servers that are DISA IA approved out of the box. Greg, thanks again for your help. We are running on a private net inside a lab, no connections outside of the lab. We'll run the NTP server either with a LOCAL reference clock driver, IRIG-B, or with GPS. GPS would be the simplest solution, and there are many classified networks with GPS timeservers, so there is ample precedent. For IA, the key is that a GPS receiver does not connect in any way to the internet, so there is no way for someone to hack in via the GPS receiver. The fact that GPS is a DoD system doesn't hurt either. A short email with Symmetricom said in essence: although there is no 'IA-mode' to put the NTP servers in, the NTP server is already running a limited amount of services, there are controls to further disable service and ports. Therefore its seems likely to me the NTP server could be configured as required. The devil is in the details however. So I would need to get funded for time to get smart on the applicable IA requirements, get a suitable COTS NTP server, configure and test it. Its likely we can get we we want, but its not going to be a simple button push like the managers would like to hear it is. Lots of things on networks lack anything resembling IA mode (whatever that is), and yet life goes on. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] IA approved COTS NTP servers question
In article 9ead1ef5-7000-445b-b7d1-ac1083874...@q8g2000vbm.googlegroups.com, Fran fran.ho...@jhuapl.edu wrote: Do you know of any DISA IA approved COTS NTP servers ? Didn¹t see any in the approved products lists at http://iase.disa.mil/common/index.html Or, have you configured/tested a COTS NTP server to pass STIG tests ? Thanks, Fran STIG: http://en.wikipedia.org/wiki/Security_Technical_Implementation_Guide I don't recall that there are any STIGs for NTP timeservers, which are based on small dedicated-mode computers running the NTP daemon under some kind of RTOS kernel. Most timeservers support at least DAC (username and password), but I don't know of any that have been evaluated to a protection profile. Which specific 8500.2 IA Controls (other than those that call out STIGs and SRGs) are you responding to? What is the threat? Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Network latency questions
In article 20100536.48397.fmgrotep...@yahoo.co.uk, Frans Grotepass fmgrotep...@yahoo.co.uk wrote: Hi all, Thanks for the responses. On Thursday 27 May 2010 15:13:50 Joseph Gwinn wrote: In article 201005271035.46352.fmgrotep...@yahoo.co.uk, Frans Grotepass fmgrotep...@yahoo.co.uk wrote: Hi all, Sorry for abusing my membership to this forum for this question. We are busy with building an embedded application that must retrieve data very fast. Please define very fast in numbers. For example, 95% of responses must be fully received within 1,000 microseconds, and 100% within 10 milliseconds, or the planet will explode. This matches the specs. Blind luck wins again. This is pretty fast for access to a remote database to work well. What does the embedded application do? SMS-routing How many subscribers? How big is the database? The choice is to either have the data locally or go to a central server(pool) that contains the data. Well, locally is always faster and more predictable than remotely, so why even consider remotely? The problem is that the remote db is already available and this will mean replicating the remote db locally. The remote db has all the data in memory. The local response time must be so fast that one needs a db solution with all the data in memory, otherwise the disk seek time will kill us. How about local caching of the database data, as needed? What fraction of the database is used locally at a worst-case instant in time? Caching can be done in local memory, using some kind of hash code access method. Local caching may be useful even if one has a local replica of the database. By the way, unless you really do need fast access to unpredictable queries, there are better designs than relational et al. Purpose-built databases often outperform general purpose databases by a factor of a hundred in speed and in footprint. In evaluating the network option, I thought that the people here could possibly help me with the expected network latency for a Gb network via a switch. My gut feeling says that with increased load, the switch will bundle the traffic to the different nodes more and this will result in higher latency. Big switches can have transit latencies of a few tens of microseconds, but there is far more to it than that. And if there is a choke point somewhere, the observed latencey will vary wildly depending on perhaps unrelated traffic and loading, making it appear that the latency varies randomly. The farther the commands and resulting data travel, the more vulnerable one is to these effects. These delays (even with local network) will make the solution impossible. Sorry again for posing the questions here. I know this is a blatant off topic post, but getting the details from the internet is a little more difficult and there are so many people here with the knowledge at hand. Thanks for the help. You're welcome. I gather you have decided that local databases are required. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Network latency questions
In article 201005271035.46352.fmgrotep...@yahoo.co.uk, Frans Grotepass fmgrotep...@yahoo.co.uk wrote: Hi all, Sorry for abusing my membership to this forum for this question. We are busy with building an embedded application that must retrieve data very fast. Please define very fast in numbers. For example, 95% of responses must be fully received within 1,000 microseconds, and 100% within 10 milliseconds, or the planet will explode. What does the embedded application do? The choice is to either have the data locally or go to a central server(pool) that contains the data. Well, locally is always faster and more predictable than remotely, so why even consider remotely? In evaluating the network option, I thought that the people here could possibly help me with the expected network latency for a Gb network via a switch. My gut feeling says that with increased load, the switch will bundle the traffic to the different nodes more and this will result in higher latency. Big switches can have transit latencies of a few tens of microseconds, but there is far more to it than that. And if there is a choke point somewhere, the observed latencey will vary wildly depending on perhaps unrelated traffic and loading, making it appear that the latency varies randomly. The farther the commands and resulting data travel, the more vulnerable one is to these effects. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] Quick sync between two computers not connected to the internet
In article 7c492e13-48c4-4b06-84ea-81e4e6596...@mac.com, Chuck Swiger cswi...@mac.com wrote: In most cases, it is easier to solve the problem of sync'ing all computers to a correct timesource (and thus all be mutually in sync), then it is to setup a bunch of truly completely isolated machines which happen to stay in sync. If I really had to solve the latter problem, I would likely connect the machines to a valid NTP timesource long enough to calibrate each machines' intrinsic drift from realtime, and then run time in standalone mode against their local clock. How good does the timekeeping need to be? Was the max error ever stated? Anyway, what I have done in such situations is to anoint a freewheeling ordinary workstation as the NTP Timeserver (trumpet flare please), and have everybody else synch to it. Synch to external time is by eyeball and wistwatch. This approach does keep them all together, but their sense of time is a good indicator of the local temperature wherever that anointed workstation is installed. That said, it is admittedly an approach taken only in desperation, and GPS-based NTP timeservers are cheap - just buy the box, install it on the network, and point everybody to it. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTPv4 Peer Event Codes - secret decoder ring sought
Dave, In article 4ba2c1ff.3060...@udel.edu, David Mills mi...@udel.edu wrote: Joe, You and Dave are working way too hard. The bits and pieces are documented on the ntpq page and on the Event Messages and Status Codes page. This would be http://www.eecis.udel.edu/~mills/ntp/html/decode.html#peer, which I didn't know about, but is exactly what I seek. And it wasn't a secret after all. But I have a question, a homework example, and a suggestion. First the question: The Code field of the Peer Status Word is 4 bits wide, and yet codes are defined for values from 1 to 10 hex (decimal 16), which doesn't quite map. How does the code value fit into the field? Wraparound, so 10 (TAI) becomes zero? The homework example: The PSW word that started this exercise is 963a. If I understand, this word decodes as follows: Status field - host_reachable plus persistent_association Select field - system_peer (gets the star) Count field - 3 Code field - become system peer (assuming code values are truncated to 4 bits, so hex 10 becomes 0) And 9614 decodes to host_reachable plus persistent_association, system_peer (gets the star), count=1, and server_reachable. And the suggestion: I was misled by some of the NTPv4 documentation, specifically the NTPv4 peerstats file documentation in http://www.eecis.udel.edu/~mills/ntp/html/monopt.html. The note under the table defining peerstats record fields reads The status field is encoded in hex format as described in Appendix B of the NTP specification RFC 1305. This is no longer really true, as you discuss below. In particular, codes exceeding 5 are not defined in 1305, and some of the definitions appear to have changed (or at least have been clarified) so it would be helpful to add a pointer to http://www.eecis.udel.edu/~mills/ntp/html/decode.html#peer to monopt.html. RFC-1305 was written in 1992. It's been 18 years since then, so you should expect changes from time to time. Changes are not done lightly; they reflect updates in the algorithms and interpretation of the statistics and state variables. If the interpretation has not changed, the name and code have not changed. If it has been changed or has become obsolete, the name is not reused. This is good. There is far too much existing base to do it any other way. Thanks, Joe Gwinn Dave Joseph Gwinn wrote: In article 46f5ae0a-93d6-44ea-812f-e4da2ae2c...@a16g2000pre.googlegroups.com, Dave Hart daveh...@gmail.com wrote: There were backward-incompatible changes on May 13, 2008 for ntp-dev 4.2.5p114: http://ntp.bkbits.net:8080/ntp-dev/?PAGE=csetREV=48295cccnu3e5cmGhOzAS7hA- pVG3A Once again statestr.c is your friend: http://ntp.bkbits.net:8080/ntp-dev/libntp/statestr.c?PAGE=diffsREV=4829513 7L4-SOuAy6YZauDbZtW6DRg If you want to be able to decode these bits for ntpd versions from before and after the change correctly, you need to query the version string of ntpd, sadly, such as with: ntpq -c rv 0 version So that's how you get the NTP version (rather than the ntpq version)! When our sysadmins first installed NTPv4, they used the version command of ntpq, which said 4. Check! I came by a few days later to look at the purported NTPv4 loopstats and peerstats files, and (ever suspicious) checked to see what version of NTP had in fact generated them. Still NTPv3. The sysadmins had been snookered by ntpq, which failed to make unambiguous whose version it was reporting upon. This had also happened to me back in the days of NTPv3, but I was saved because I knew that 4 could not be the answer. But I never did figure out how to get ntpq to tell me the version of the ntp daemon. and then parse for 4.2.5p114 or later. The format for the version string can include an optional -RC# suffix, and before long, there may be releases with a -beta# suffix in the -stable branch, such as 4.2.6p2-beta1 as a prelude to 4.2.6p2-RC1. Still evolving, rapidly. OK. I will have to find out exactly which version I have. I have no need to decode status from prior versions. I need only to understand the status codes from what I am running, to understand what is and is not working in my system. Fixes have included giving NTP and related traffic its own dedicated LAN and LAN ports on the hosts, to reduce buffeting of NTP packets and/or the daemon by unrelated but heavy packet traffic. The buffeting causes what appear to be large, random, and often asymmetric transport delays. Is there available a written discussion of which changes were made and why? This could be worth reading. More generally, these backward-incompatible changes will cause great confusion and difficulty in transitioning to NTPv4 unless ntpq is kept up to date, and the descriptions of what the various status codes mean are both complete and correct - telegraphic summaries
Re: [ntp:questions] NTPv4 Peer Event Codes - secret decoder ring sought
Dave, In article 81ed5f77-97a2-474d-8c1a-346b2192c...@v34g2000prm.googlegroups.com, Dave Hart daveh...@gmail.com wrote: On Mar 18, 13:49 UTC, Joseph Gwinn joegw...@comcast.net wrote: Dave Hart daveh...@gmail.com wrote: If you want to be able to decode these bits for ntpd versions from before and after the change correctly, you need to query the version string of ntpd, sadly, such as with: ntpq -c rv 0 version So that's how you get the NTP version (rather than the ntpq version)! When our sysadmins first installed NTPv4, they used the version command of ntpq, which said 4. Check! I came by a few days later to look at the purported NTPv4 loopstats and peerstats files, and (ever suspicious) checked to see what version of NTP had infact generated them. Still NTPv3. The sysadmins had been snookered by ntpq, which failed to make unambiguous whose version it was reporting upon. This had also happened to me back in the days of NTPv3, but I was saved because I knew that 4 could not be the answer. But I never did figure out how to get ntpq to tell me the version of the ntp daemon. C:\NTPb\binntpq --version ntpq - standard NTP query program - Ver. 4.2.7p20 C:\NTPb\binntpq -c version ntpq 4.2.7...@1.2137-o Mar 18 15:04:17.18 (UTC-00:00) 2010 (4) C:\NTPb\binntpq -c rv 0 version version=ntpd 4.2.7...@1.2137-o Mar 14 8:23:33.64 (UTC-00:00) 2010 (9) C:\NTPb\bin The first two commands above are both reporting on the ntpq version, in slightly different form. The third reports on the local ntpd version. Tack on a hostname or IP address, and it'll tell you about a remote ntpd version, if you're allowed to use ntpq with the server in question. This is a very useful summary. I'll pass it on to the sysadmins. Is there available a written discussion of which changes were made and why? This could be worth reading. If there is, it would be in the archives of committers@, hackers@, or questions@lists.ntp.org (all browsable via http://lists.ntp.org/) from around May 13, 2008. I was not active on the lists at that time. I'll poke around. It will no doubt help understanding the genesis of the status codes and their descriptions in http://www.eecis.udel.edu/~mills/ntp/html/decode.html#peer. Looking at the code you suggested, I also see that the variable names are the same as in NTPv3 (and the names imply the original NTPv3 meanings), but the new NTPv4 comments on those variables seem to contradict the meanings implied by the names. Not knowing the history makes it difficult to figure out just what is now meant. I believe the 2008 changes were part of overall cleanup to bring the reference implementation in-line with the draft NTP v4 specification. The RFC form of that document has just been approved by the IESG and should be a proposed standard RFC before too many more weeks. Please refer to that document in your search for meaning: ftp://ftp.rfc-editor.org/in-notes/internet-drafts/draft-ietf-ntp-ntpv4-proto-13.txt Which is derived from the less ASCII-hamstrung: http://www.eecis.udel.edu/~mills/database/reports/ntp4/ntp4.pdf I have read ntp4.pdf (dated June 2006), and it says nothing of status codes and the like. I assume that this is intentional, and that one is expected to consult the online documentation for such information. Perhaps the text ties to the codes in decode.html#peer. Thanks, Joe ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTPv4 Peer Event Codes - secret decoder ring sought
In article slrnhq850s.183.un...@wormhole.physics.ubc.ca, unruh un...@wormhole.physics.ubc.ca wrote: On 2010-03-19, David Mills mi...@udel.edu wrote: Joe, That's a typo; event 16 does not exist. Glad you caught that. Pretty elaborate typo. Did they mean to give it a number other than 16, or were 50 letters somehow mistyped? Ahh, be nice. We all know perfectly well how such things happen. Joe Gwinn Joseph Gwinn wrote: Dave, In article 4ba2c1ff.3060...@udel.edu, David Mills mi...@udel.edu wrote: Joe, You and Dave are working way too hard. The bits and pieces are documented on the ntpq page and on the Event Messages and Status Codes page. This would be http://www.eecis.udel.edu/~mills/ntp/html/decode.html#peer, which I didn't know about, but is exactly what I seek. And it wasn't a secret after all. But I have a question, a homework example, and a suggestion. First the question: The Code field of the Peer Status Word is 4 bits wide, and yet codes are defined for values from 1 to 10 hex (decimal 16), which doesn't quite map. How does the code value fit into the field? Wraparound, so 10 (TAI) becomes zero? The homework example: The PSW word that started this exercise is 963a. If I understand, this word decodes as follows: Status field - host_reachable plus persistent_association Select field - system_peer (gets the star) Count field - 3 Code field - become system peer (assuming code values are truncated to 4 bits, so hex 10 becomes 0) And 9614 decodes to host_reachable plus persistent_association, system_peer (gets the star), count=1, and server_reachable. And the suggestion: I was misled by some of the NTPv4 documentation, specifically the NTPv4 peerstats file documentation in http://www.eecis.udel.edu/~mills/ntp/html/monopt.html. The note under the table defining peerstats record fields reads The status field is encoded in hex format as described in Appendix B of the NTP specification RFC 1305. This is no longer really true, as you discuss below. In particular, codes exceeding 5 are not defined in 1305, and some of the definitions appear to have changed (or at least have been clarified) so it would be helpful to add a pointer to http://www.eecis.udel.edu/~mills/ntp/html/decode.html#peer to monopt.html. RFC-1305 was written in 1992. It's been 18 years since then, so you should expect changes from time to time. Changes are not done lightly; they reflect updates in the algorithms and interpretation of the statistics and state variables. If the interpretation has not changed, the name and code have not changed. If it has been changed or has become obsolete, the name is not reused. This is good. There is far too much existing base to do it any other way. Thanks, Joe Gwinn Dave Joseph Gwinn wrote: In article 46f5ae0a-93d6-44ea-812f-e4da2ae2c...@a16g2000pre.googlegroups.com, Dave Hart daveh...@gmail.com wrote: There were backward-incompatible changes on May 13, 2008 for ntp-dev 4.2.5p114: http://ntp.bkbits.net:8080/ntp-dev/?PAGE=csetREV=48295cccnu3e5cmGhOzAS7 hA- pVG3A Once again statestr.c is your friend: http://ntp.bkbits.net:8080/ntp-dev/libntp/statestr.c?PAGE=diffsREV=4829 513 7L4-SOuAy6YZauDbZtW6DRg If you want to be able to decode these bits for ntpd versions from before and after the change correctly, you need to query the version string of ntpd, sadly, such as with: ntpq -c rv 0 version So that's how you get the NTP version (rather than the ntpq version)! When our sysadmins first installed NTPv4, they used the version command of ntpq, which said 4. Check! I came by a few days later to look at the purported NTPv4 loopstats and peerstats files, and (ever suspicious) checked to see what version of NTP had in fact generated them. Still NTPv3. The sysadmins had been snookered by ntpq, which failed to make unambiguous whose version it was reporting upon. This had also happened to me back in the days of NTPv3, but I was saved because I knew that 4 could not be the answer. But I never did figure out how to get ntpq to tell me the version of the ntp daemon. and then parse for 4.2.5p114 or later. The format for the version string can include an optional -RC# suffix, and before long, there may be releases with a -beta# suffix in the -stable branch, such as 4.2.6p2-beta1 as a prelude to 4.2.6p2-RC1. Still evolving, rapidly. OK. I will have to find out exactly which version I have. I have no need to decode status from prior versions. I need only to understand the status codes from what I am running, to understand what is and is not working in my system. Fixes have included giving NTP and related traffic its own dedicated LAN and LAN ports
Re: [ntp:questions] NTPv4 Peer Event Codes - secret decoder ring sought
In article 46f5ae0a-93d6-44ea-812f-e4da2ae2c...@a16g2000pre.googlegroups.com, Dave Hart daveh...@gmail.com wrote: There were backward-incompatible changes on May 13, 2008 for ntp-dev 4.2.5p114: http://ntp.bkbits.net:8080/ntp-dev/?PAGE=csetREV=48295cccnu3e5cmGhOzAS7hA-pVG3A Once again statestr.c is your friend: http://ntp.bkbits.net:8080/ntp-dev/libntp/statestr.c?PAGE=diffsREV=48295137L4-SOuAy6YZauDbZtW6DRg If you want to be able to decode these bits for ntpd versions from before and after the change correctly, you need to query the version string of ntpd, sadly, such as with: ntpq -c rv 0 version So that's how you get the NTP version (rather than the ntpq version)! When our sysadmins first installed NTPv4, they used the version command of ntpq, which said 4. Check! I came by a few days later to look at the purported NTPv4 loopstats and peerstats files, and (ever suspicious) checked to see what version of NTP had in fact generated them. Still NTPv3. The sysadmins had been snookered by ntpq, which failed to make unambiguous whose version it was reporting upon. This had also happened to me back in the days of NTPv3, but I was saved because I knew that 4 could not be the answer. But I never did figure out how to get ntpq to tell me the version of the ntp daemon. and then parse for 4.2.5p114 or later. The format for the version string can include an optional -RC# suffix, and before long, there may be releases with a -beta# suffix in the -stable branch, such as 4.2.6p2-beta1 as a prelude to 4.2.6p2-RC1. Still evolving, rapidly. OK. I will have to find out exactly which version I have. I have no need to decode status from prior versions. I need only to understand the status codes from what I am running, to understand what is and is not working in my system. Fixes have included giving NTP and related traffic its own dedicated LAN and LAN ports on the hosts, to reduce buffeting of NTP packets and/or the daemon by unrelated but heavy packet traffic. The buffeting causes what appear to be large, random, and often asymmetric transport delays. Is there available a written discussion of which changes were made and why? This could be worth reading. More generally, these backward-incompatible changes will cause great confusion and difficulty in transitioning to NTPv4 unless ntpq is kept up to date, and the descriptions of what the various status codes mean are both complete and correct - telegraphic summaries are not usually enough for non-developers to understand. Looking at the code you suggested, I also see that the variable names are the same as in NTPv3 (and the names imply the original NTPv3 meanings), but the new NTPv4 comments on those variables seem to contradict the meanings implied by the names. Not knowing the history makes it difficult to figure out just what is now meant. Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] NTPv4 Peer Event Codes - secret decoder ring sought
In article joegwinn-1c741d.09115717032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net wrote: Dave, In article 8c5b8d60-8780-4bf0-80da-6b1d19410...@k24g2000pro.googlegroups.com, Dave Hart daveh...@gmail.com wrote: On Mar 17, 03:30 UTC, Joseph Gwinn wrote: Looking in section B.2.2 of RFC 1305 yields that the Peer Status Field has four subfields, the last (rightmost) one of which being the 4-bit Peer Event Code (page 57), which is defined for values between 0 and 5, and is reserved for values 6 to 15. Well, I have been seeing two values of Peer Status, 9614 and 963a, both hexidecimal. I understand 9614, but 963A is a mystery, as it implies a Peer Event Code of 10 (the A in the rightmost digit), which is undefined and reserved in RFC 1305. Scan for PEVNT_ in ntp.h: http://ntp.bkbits.net:8080/ntp-stable/include/ntp.h?PAGE=annoREV=4af5f8cfD BBhNWjyJ4XiD74vlioxeg #define PEVNT_MOBIL (1 | PEER_EVENT) /* mobilize */ #define PEVNT_DEMOBIL (2 | PEER_EVENT) /* demobilize */ #define PEVNT_UNREACH (3 | PEER_EVENT) /* unreachable */ #define PEVNT_REACH (4 | PEER_EVENT) /* reachable */ #define PEVNT_RESTART (5 | PEER_EVENT) /* restart */ #define PEVNT_REPLY (6 | PEER_EVENT) /* no reply */ #define PEVNT_RATE (7 | PEER_EVENT) /* rate exceeded */ #define PEVNT_DENY (8 | PEER_EVENT) /* access denied */ #define PEVNT_ARMED (9 | PEER_EVENT) /* leap armed */ #define PEVNT_NEWPEER (10 | PEER_EVENT) /* sys peer */ #define PEVNT_CLOCK (11 | PEER_EVENT) /* clock event */ #define PEVNT_AUTH (12 | PEER_EVENT) /* bad auth */ #define PEVNT_POPCORN (13 | PEER_EVENT) /* popcorn */ #define PEVNT_XLEAVE(14 | PEER_EVENT) /* interleave mode */ #define PEVNT_XERR (15 | PEER_EVENT) /* interleave error */ #define PEVNT_TAI (16 | PEER_EVENT) /* TAI */ This almost answers the immediate question, but what exactly is a sys peer event? More generally, the comments on many event codes lack verbs, defeating the reader. Perhaps these codes are expanded in ntpq; I'll look ... To match the literal text output by ntpd/ntpq when decoding, see also libntp/statestr.c: http://ntp.bkbits.net:8080/ntp-stable/libntp/statestr.c?PAGE=annoREV=4ac6e 036jH41_maMfVXyf2VeiFknzQ ... by following this breadcrumb trail. There may be an easier way, but looking at the source comes naturally to me. There isn't an easier way for most people until the NTPv4 documentation is updated, which is essential. Right now, the NTPv4 documentation points users and sysadmins to an authoritative but incomplete answer. I knew that the answer had to be in the ~70,000 lines of NTP source code, but wouldn't really know which rock to look under. Very few people have the time to know this much source code well enough to find the correct answer, and to know that the found answer is in fact correct. Which is why the NTPv4 documentation needs to be revised to reflect the as-built NTPv4 code. NTP has many millions of users, but at most a few hundred developers (where a developer is someone who knows his way around the source code). Well, I started writing the decoder for NTPv4 Peer Status, and soon fetched up on the rocks. It appears from the names that the definitions of the status bits have changed. I recall someone saying on comp.protocols.time.ntp that this was the case, and now I see what books like confirmation. Things have changed, and yet one can convince oneself that these old and new variables are actually the same. Maybe it's really true. Another issue is that the other fields of the Peer Status Word may have changed. Which C structures correspond to which tables in the online documentation? I will need enough structure that I can hand-decode an arbitrarily built but compliant Peer Status Word. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
[ntp:questions] NTPv4 Peer Event Codes - secret decoder ring sought
Well, we just brought NTPv4 up on some IBM AIX 5.3 machines. Had to compile from source code on the target machines to get a daemon that didn't crash upon launch. Anyway, the daemon appears to be happily working, and is happily generating loopstats and peerstats files. So far so good. The peerstats files contain a Peer Status Word field. NTPv4 is documented in http://www.eecis.udel.edu/~mills/ntp/html/index.html, and NTPv4 peerstats files are documented in http://www.eecis.udel.edu/~mills/ntp/html/monopt.html. The note under the table defining peerstats record fields reads The status field is encoded in hex format as described in Appendix B of the NTP specification RFC 1305. (The draft RFC for NTPv4 is innocent of all such status information.) Looking in section B.2.2 of RFC 1305 yields that the Peer Status Field has four subfields, the last (rightmost) one of which being the 4-bit Peer Event Code (page 57), which is defined for values between 0 and 5, and is reserved for values 6 to 15. Well, I have been seeing two values of Peer Status, 9614 and 963a, both hexidecimal. I understand 9614, but 963A is a mystery, as it implies a Peer Event Code of 10 (the A in the rightmost digit), which is undefined and reserved in RFC 1305. I would guess that NTPv4 has used some of the codes that were held in reserve in NTPv4, but where are these new codes formally defined? This is most likely a general question, and I would hazard that this isn't the only place where NTPv4 has outrun its documentation. Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] National time standard differences
In article c2bcn.37983$ym4.23...@text.news.virginmedia.com, David J Taylor david-tay...@blueyonder.delete-this-bit.and-this-part.co.uk.invalid wrote: I've setup an NTP server in the south east asia region synchonising with regional NTP servers as well as a couple of servers I am responsible for in the US. remotest t when poll reach delay offset jitter == +Japan 1 u 454 1024 377 87.402 -0.560 0.056 *Japan 1 u 476 1024 377 87.277 -0.810 1.542 -NorthAmerica 2 u 427 1024 377 285.387 -17.741 5.084 -NorthAmerica 2 u 429 1024 377 307.208 -17.061 0.083 Is the above delta seen between Japanese and North American time sources purely the delta between the outbound / return network path or something else? Chris Chris, Most likely asymmetrical paths, yes. The wide area networks used for international connections are typically SONET rings, which are inherently asymmetrical. As one works around a ring, the poll response time is constant (being the perimeter of the ring), while the out and back times vary with position on the ring with respect to the server one is polling. Joe Gwinn ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] 500ppm - is it too small?
In article lak0t6-g6n@klein-habertwedt.de, Uwe Klein uwe_klein_habertw...@t-online.de wrote: Joseph Gwinn wrote: The prototypical example of an orthogonal instruction set was the PDP-11. The Motorola 68000 family was an outgrowth. 68k - CISC and still very much alive 88k - RISC drifting belly up in the pond. in ~1990 Motorola wanted to push RISC 88k so much you could buy a MVME187 plus SysVR3 unix System for half the price of a MVME167. Afaik it was a flop. I remember the 88K. Sort of. PowerPC later took off in an acceptable way. Yes. IBM uses the instruction set to this day in their large servers. Another side is a bit more interesting: Motorola had a perfect design for the 68k processor developement path at the time they released the inital MC68000. Things just worked. Yes. I did a fair bit of 680x0 programming over the years. The 680x0 still lives, having become a tiny little embedded computer chip for use in appliance controllers and the like. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] 500ppm - is it too small?
In article 87r5s3syxz@pc9454.klinik.uni-regensburg.de, Ulrich Windl ulrich.wi...@rz.uni-regensburg.de wrote: nemo_outis a...@xyz.com writes: Richard B. Gilbert rgilber...@comcast.net wrote in news:tlsdnq2e26bblbnxnz2dnuvz_sydn...@giganews.com: nemo_outis wrote: ... I fail to see the value or relevance of 500ppm satisfies 98% of computer clocks if some other number, perhaps 5000 ppm, could satisfy yet even more than 98% of computer clocks with no downside - as indeed seems to be the case! Chrony, whatever its other merits and demerits, is an existence proof for this proposition. I can't follow Dave's math but I'm reasonably sure that there is a good reason for the 500 PPM limit. Since almost all computer clocks can meet this criterion I'm not going to worry about it. Hmm, faith-based ntp? Not for me. If there is a good reason I'd like to hear it - 500 ppm has the smell of arbitrariness about it. As arbitrary as there are 8 bits in a byte. No, 8 bits isn't arbitrary. Computer hardware is simplified if the various word lengths are all powers of two. Eight bits was the smallest power-of-two size that allowed the full Roman alphabet including punctuation and control characters to be coded. There are 5, 6, and 7 bit codes, all now obsolete: Five-bit: Baudot, used in teletypes. Six-bit: Fieldata (Univac and Control Data, and others I assume.) Seven-bit: ASCII without parity bit. Eight bit: ASCII with parity bit, and EBCDIC (http://www.ncsa.illinois.edu/UserInfo/Resources/Hardware/IBMp690/IBM/usr /share/man/info/en_US/xlf/html/lr425.HTM) ASCII came from ATT, while EBCDIC came from IBM. And now sixteen bit: Unicode. (http://unicode.org/standard/WhatIsUnicode.html) Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] .1 Microsecond Synchronization
In article 2pmdnygdgrswzbrxnz2dnuvz_uidn...@giganews.com, ScottyG sco...@pepex.net wrote: Hello. The company I am working for needs to be able to record timestamps in a trading system logs down to a .1 microsecond accuracy. We will have servers located in London, New York and Chicago. There will be a dedicated resilient link between London and New York. Searches on the web have made claims that NTP can achieve this accuracy. Unfortunately the sales rep for the NTP server we looked at told me that the best I could expect is 2-5 ms synchronization across servers. Has anyone had any experience doing this? Can anyone suggest how to achieve this accuracy? We do have some budget but this but if I need to spend a whole lot on this I need to get in front of my management with the reasons. Thank you in advance for any help or suggestions you can give me. It's not obvious that what the company wants to do is even physically possible. As many others have mentioned, getting 100 nanosecond synch everywhere in a network 10,000 km in diameter isn't easy, and isn't possible without expensive specialized hardware. However, even with the special hardware, there are very deep problems with the whole idea. The classic reference is: Time, clocks, and the ordering of events in a distributed system, Leslie Lamport, Communications of the ACM 21,7 (July 1978) pages 558-565. Paper 26 in http://research.microsoft.com/en-us/um/people/lamport/pubs/pubs.html. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article ywn94oxs3u9t@ntp1.isc.org, Harlan Stenn st...@ntp.org wrote: In article joegwinn-9c92e9.21153416032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net writes: Joseph In article ywn9tz5tflz6@ntp1.isc.org, Joseph Harlan Stenn st...@ntp.org wrote: In article joegwinn-6fd03a.17481615032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net writes: I think you are talking about one of my pet peeves: http://support.ntp.org/bin/view/Dev/NtpVariablesAndNtpq Joseph I don't think that I have inconsistent versions of ntpd and ntpq, Joseph because both came off the same CD from Sun Microsystems. It's still the same beast. The bottom line is we currently have opaque data being presented to the user, and that is either being offered directly to the user (in your case) or is being potentially mis-converted by ntpq. Joseph I have a lot of trouble believing that Sun put inconsistent versions Joseph on their Solaris install CDs. I was not talking about the inconsistent version problem. I'm talking about opaque data. Joseph Nor am I using NTPQ for decoding. I decode these codes myself, Joseph following Appendix B of RFC-1305. It turns out that NTPv4 uses the Joseph same definitions. See Joseph http://www.eecis.udel.edu/~mills/ntp/html/monopt.html. Then I may have misunderstood. My point is that while it's fine for ntpd to send encoded data, I think we need to have a way for that data to *also* be sent decoded, or provide enough information so programs like ntpq can decode the result, regardless of which version of ntpd they are talking to. It isn't quite bulletproof, but my decoder code also tells NTPv3 and NTPv4 loopstats and peerstats records apart, keyed on (loopstats?) record length. One can only hope that NTPv5 et seq will ensure that that an unsupervised and simple decoder program is able to tell record formats from various NTP versions apart. As for a decoded format, that is pure ascii, that would certainly be the unix way, and would impose negligible load on all but the most skeletal of embedded systems. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
[ntp:questions] NTP Support (Was 'What does Max Distance Exceeded...')
We have moved from the meaning of status code 9514 to the more general issue of how NTP shall be supported, so I've collected the relevant threads below. === At 11:19 PM -0400 3/15/09, Danny Mayer wrote: Joseph Gwinn wrote: The FAQ has to be the place for such explanations. I'm not sure if this qualifies as an FAQ as I don't recall that it has come up before. FAQ stands for Frequently Asked Questions. RAQ then? Rarely Asked Questions Seriously, I can't believe that I'm the only person in history to be perplexed by these status codes, and those little three-word summaries are a bit telegraphic. Joe Gwinn You aren't the only one. These questions have been asked before by a number of people. In fact I had to look at this at one point when I was getting these codes. Of course I just looked at the source code and never looked for documentation. I will tell you that this is a combination of bits so it's not just a number. Each bit represents a test code that failed so you have quite a bit to look at. I do know how the status code is structured, and wrote a Mathematica program to automate the decoding. (I use Mathematica to generate the co-plots of loopstats and peerstats data, collect statistics, et al.) What I didn't know was that the definitions of the code bits had changed between v3 and v4. I'll have to dig into the old documentation and see if this code was affected. There is little chance that I will have the time to read enough NTP source code to make sense of it, sufficient to be able to come to reliable conclusions. I'm a system engineer, and time is one issue of many in a system. More generally, it's hopeless to expect the world's sysadmins to read NTP code (or any other kind of code). They just don't have the time, and are responsible for far too many different kinds of box for it to be practical. But a major part of making something reliable in practice is making it possible for a harried sysadmin to nonetheless get it right. (I'm not a sysadmin, but work with many sysadmins. They spend lots of time fighting fires, and are of necessity jacks of all trades, masters of none.) Silently mutating code definitions sounds like a blunder to me. NTP is used on tens to hundreds of millions of computers worldwide. There will never be a pure v4 world. In fact there will still be v3 around when v5 is being introduced. So, if new kinds of status is needed, invent new codes to suit, but do not change the meanings of the codes that are already widely used. In other words, do not undermine your existing base. The Internet folk had the same issue with IPv6, and they concluded that IPv4 was too deeply embedded to ever eliminate, and that there was never going to be a flag day when a worldwide changeover would happen. Thus, IPv4 and IPv6 had to coexist and interoperate forever, and so IPv6 was designed to support this. == To: ma...@ntp.org From: Joe Gwinn joegw...@comcast.net Subject: Re: [ntp:questions] What exactly does Maximum Distance Exceded mean? Cc: questions@lists.ntp.org Bcc: gw...@raytheon.com X-Attachments: Status code values fixed. At 10:47 PM -0400 3/15/09, Danny Mayer wrote: Joseph Gwinn wrote: Hmm. OK, but I think that we've kind of run off the rails. Let me summarize: 1. Sun Microsystems' current behavior is not the issue, as I'm loading old software from an old CD onto old computer hardware, hardware that cannot support a newer version of Solaris than v9. One of these old Solaris boxes did work with NTPv3 running an even older version of Solaris, with no 9514 codes, deepening the mystery. The trouble here is that those codes are *very likely* likely to have changed between V3 and V4 since there was a large rewrite between the two. That's why looking at the source code is necessary to get you the help you need. As discussed in my other reply, mutating codes is a blunder. It's a good-news bad-news thing. The good news is that NTP has succeeded on an unimagined scale. The bad news is that because of that scale, one must be *very* respectful of NTP's existing base, and it *can* be constraining. The fact that this obsolete system can most likely support NTPv4 is worth investigation, though. 2. I think that what's happening is that I'm doing something dumb, and I bet that there is no real difference in how NTPv3 or NTPv4 would react to this faux pas, whatever it turns out to be. Nor is source code research needed or requested. 3. The original question was how to interpret a specific status code, 9514. I read the explanation in the documentation, but became no wiser for it. Thus my question. Which is why you need to look at the source code. Documentation isn't always clear or definitive but the source code will tell you
Re: [ntp:questions] NTP Support (Was 'What does Max Distance Exceeded...')
In article 49becf09$0$507$5a6ae...@news.aaisp.net.uk, David Woolley da...@ex.djwhome.demon.co.uk.invalid wrote: Joseph Gwinn wrote: We have moved from the meaning of status code 9514 to the more general But you should have kept the thread, even if the subject changed. Opinion varies on this, but why? It really was a case of topic drift. issue of how NTP shall be supported, so I've collected the relevant threads below. More generally, it's hopeless to expect the world's sysadmins to read NTP code (or any other kind of code). They just don't have the time, Generally, you only need to read a small bit of code to answer this sort of question, but if you haven't got the time you should pay someone who does have the time. It's simply not going to happen, especially for random small questions. They will just muddle through, and blame NTP. Hiring outside help isn't just a money problem, it also requires much jumping through bureaucratic hoops, which is very time consuming, so it never makes sense for non-major issues. Historically, open source software was written for use by people who had the ability to support it themselves. Recently, the relationship has become asymmetric with a lot of people wanting free software and free support. Whilst some open source software developers may consider it a valuable loss leader to produce a naive user product and support it, may even consider it part of their mission, most open source developers are not that interested in donating that level of free support. What price success? Most open-source software would be happy to achieve 1% of what NTP has achieved. Only Linux is even in the running, but NTP far exceeds Linux. Given the present scale, for which NTP and its community were never designed, what to do? The single most effective thing the community can do is to write good documentation. Yes, it's work, but it's by far the most effective thing one can do, and it's really the only practical approach given the immense size of the NTP user base. By the way, when I started digging into the archived NTPv3 documentation, it said that the peerstats status codes were defined in Appendix B of RFC-1305. OK, I actually knew that. Then, I started looking for the corresponding NTPv4 documentation. The NTPv4 RFC to be is innocent of the word peerstats and its status field, but the current online documentation (which is for NTPv4) at http://www.eecis.udel.edu/~mills/ntp/html/monopt.html also points to the same place in RFC-1305, so it appears that the status code definitions did not change. Unless the documentation is in error. If so, the simplest global fix is to update the documentation. As I said in the prior posting, it's a non-starter to try to require a hundred million people to either read the source code or pay someone to do it for them every time NTP throws an uncommon status code. It just won't happen. On scale alone, the NTP community would be overwhelmed with repeated trivial questions. I'm reminded of the librarians in the town library when I was in high school. There were lots of hand-written cards in the card file, placed there by the librarians when they got tired of hearing the same simple or silly question. Actually a lot of commercial software, these days, is dumbed down, supported, open source material. True. But it can be amusing to figure out which one it is, and start to ask uncomfortable questions about how can they be better than free, especially if they reduced the usefulness of the code. Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 20090316063319.f2...@bulldog.localhost.org, hun...@comcast.net (Rob Neal) wrote: On Sun, 15 Mar 2009, Joseph Gwinn wrote: In article qo6dnzljdv8z4cdunz2dnuvz_hwwn...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: [snip] The FAQ has to be the place for such explanations. I'm not sure if this qualifies as an FAQ as I don't recall that it has come up before. FAQ stands for Frequently Asked Questions. RAQ then? Rarely Asked Questions Seriously, I can't believe that I'm the only person in history to be perplexed by these status codes, and those little three-word summaries are a bit telegraphic. You have lots of company, sadly. A decoder function would be 'Good'. There are obstacles to this, as remarked by others, and a serious lack of volunteer time to code a solution. No one will stop you, if you wish to contribute In the current build I find the TEST status codes in ntp.h. They have changed, from release to release, so consult your source for particulars. There has been considerable improvement in NTP from V3 to V4. You should consider upgrading, it really is better. I already have a decoder function, coded in Mathematica, based on Appendix B of RFC-1305, which ought to work for NTPv3. The problem is one of documentation, as RFC-1305 is pretty telegraphic, although it does point one to the descriptions of the relevant tests. I gather that the actual NTPv3 code does not really follow RFC-1305. That seems to be the bottom line. Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 1237210896.237...@news1nwk, Brian Utterback brian.utterb...@sun.com wrote: Joseph Gwinn wrote: Also good to know, so I'll know better than to use the Sun compiler (not that it's bad, but that it isn't what's been worked through with NTP). Not to worry, I make sure that the current -dev branch will always build with the Sun Studio compilers. Ahh. Good to know. Thanks, Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article ywn9y6v5fm70@ntp1.isc.org, Harlan Stenn st...@ntp.org wrote: In article joegwinn-4a6c5f.16502115032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net writes: Joseph What AIX version and Technology Level (~=patch level) have been used Joseph to build NTPv4? powerpc-ibm-aix4.3.3.0 powerpc-ibm-aix5.1.0.0 powerpc-ibm-aix5.2.0.0 powerpc-ibm-aix5.3.0.0 Thanks. We are using AIX 5.3 TL5 and TL6. Although there is loose talk of going to AIX 6.1, who knows if this will soon happen. Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article zcsdncxjv-jn2spunz2dnuvz_gawn...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article qo6dnzljdv8z4cdunz2dnuvz_hwwn...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article nbwdnvxq_p-y_idunz2dnuvz_thin...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article 49bd3907.1080...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: [snip] 3. The original question was how to interpret a specific status code, 9514. I read the explanation in the documentation, but became no wiser for it. Thus my question. If there isn't a NTP FAQ entry on this, there probably should be. Our sysadmins were flummoxed by the cloud of 5914 codes, and they are far too busy to undertake a research project. (The deeper problem is that some managers believe that NTP is plug and play, which isn't quite true.) The various answers and questions I've gotten have been quite useful, as they give me a list of things to think about and investigate, things I might not have thought of, or soon thought of. Joe Gwinn Joe, You need to proofread your message text a little more carefully!! Which error are you ACTUALLY getting? You say 9514 and then 5914! Which is it? You're right, but it wouldn't help, for an odd reason. The status code is 9514. But I have a Clausing 5914 lathe. Inherent dyslexia inducer. Also, you might try Google with the FULL and EXACT text of the error message! It's 9514, pulled from a field in peerstats records. Think I'll get many false hits? Qualifying 9514 with peerstats brought me back to this thread. So, tried Maximum Distance Exceded, got led back to this exact news thread. But let's say I did find some relevant hits. This is the Internet. How would I know which hits to believe? I would be influenced by who wrote it and who disagreed with him! But what if I listen to the loudest one? The FAQ has to be the place for such explanations. I'm not sure if this qualifies as an FAQ as I don't recall that it has come up before. FAQ stands for Frequently Asked Questions. RAQ then? Rarely Asked Questions Seriously, I can't believe that I'm the only person in history to be perplexed by these status codes, and those little three-word summaries are a bit telegraphic. Joe Gwinn I don't see any of the usual suspects stepping up to take responsibility. You may have to reverse engineer the code in order to satisfy your curiosity. Actually, in the earlier part of this thread, better explanations were given, before we drifted off into NTP support models. Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article ywn9tz5tflz6@ntp1.isc.org, Harlan Stenn st...@ntp.org wrote: In article joegwinn-6fd03a.17481615032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net writes: I think you are talking about one of my pet peeves: http://support.ntp.org/bin/view/Dev/NtpVariablesAndNtpq Joseph I don't think that I have inconsistent versions of ntpd and ntpq, Joseph because both came off the same CD from Sun Microsystems. It's still the same beast. The bottom line is we currently have opaque data being presented to the user, and that is either being offered directly to the user (in your case) or is being potentially mis-converted by ntpq. I have a lot of trouble believing that Sun put inconsistent versions on their Solaris install CDs. Nor am I using NTPQ for decoding. I decode these codes myself, following Appendix B of RFC-1305. It turns out that NTPv4 uses the same definitions. See http://www.eecis.udel.edu/~mills/ntp/html/monopt.html. Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49be52ab.1000...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Joe Gwinn wrote: Status code values fixed. At 10:47 PM -0400 3/15/09, Danny Mayer wrote: Joseph Gwinn wrote: Hmm. OK, but I think that we've kind of run off the rails. Let me summarize: 1. Sun Microsystems' current behavior is not the issue, as I'm loading old software from an old CD onto old computer hardware, hardware that cannot support a newer version of Solaris than v9. One of these old Solaris boxes did work with NTPv3 running an even older version of Solaris, with no 9514 codes, deepening the mystery. The trouble here is that those codes are *very likely* likely to have changed between V3 and V4 since there was a large rewrite between the two. That's why looking at the source code is necessary to get you the help you need. As discussed in my other reply, mutating codes is a blunder. It's a good news bad news thing. The good news is that NTP has succeeded on an unimagined scale. The bad news is that because of that scale, one must be *very* respectful of NTP's existing base, and it can be constraining. You won't get any argument from us. However, Dave Mills is responsible for these codes and we haven't been able to get him to agree to not change the test code numbers and to use new ones if he needs more and just not reuse the old ones. He has good reasons for changing the tests but changing the meaning of the same code is harder to fathom. His view is that these are internal tests but when you are trying to track down a problem with your ntp daemon, it's important to know what they mean. These codes are *not* internal only. They are documented in RFC-1305, Appendix B, which is also pointed to by the NTPv4 documentation http://www.eecis.udel.edu/~mills/ntp/html/monopt.html. These codes are quite public. The fact that this obsolete system can most likely support NTPv4 is worth investigation, though. 2. I think that what's happening is that I'm doing something dumb, and I bet that there is no real difference in how NTPv3 or NTPv4 would react to this faux pas, whatever it turns out to be. Nor is source code research needed or requested. 3. The original question was how to interpret a specific status code, 9514. I read the explanation in the documentation, but became no wiser for it. Thus my question. Which is why you need to look at the source code. Documentation isn't always clear or definitive but the source code will tell you. It simply cannot be required to read source code to get the definitions of status codes, even if the documentation has to give one definition per NTP version. NTP is used on hundreds of millions of computers. Are we expecting that every time someone gets an unexpected code they either have to read the source code, or pay someone to read it for them? I'm sorry, but that cannot work. I agree, but I'm not the person you need to persuade. In V4 the flash codes are listed in libntp/statestr.c. I don't know about V3. While given the pointer I may well look, the fundamental issue remains. You may also be amused by this sync code: { CTL_SST_TS_WRSTWTCH, sync_wristwatch }, Heh. Hairy wrist required. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49bdbbbe.4030...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Joseph Gwinn wrote: What's the story for IBM's AIX? It builds on AIX too. It builds on most Unix systems though maybe not on some of the oldest O/S versions. Including the AIX we use, as mentioned in another posting. Joe Gwinn PS: I'll be offline for the next two weeks, starting tomorrow morning. ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 89b3231f-d79a-4fb1-b449-fe574bca8...@j38g2000yqa.googlegroups.com, paul.cro...@softwareag.com wrote: Joseph, If you're not willing to get the source code for NTP and compile it, you can download a binary from http://www.sunfreeware.com/. It's probably configured with a 'standard' set of refclock drivers and, as a consequence, may be larger than a custom-configured version. We will most likely do just that. Size isn't much of a problem, especially for an experiment. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49bdc52b.20...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Joseph Gwinn wrote: The FAQ has to be the place for such explanations. I'm not sure if this qualifies as an FAQ as I don't recall that it has come up before. FAQ stands for Frequently Asked Questions. RAQ then? Rarely Asked Questions Seriously, I can't believe that I'm the only person in history to be perplexed by these status codes, and those little three-word summaries are a bit telegraphic. Joe Gwinn You aren't the only one. These questions have been asked before by a number of people. In fact I had to look at this at one point when I was getting these codes. Of course I just looked at the source code and never looked for documentation. My fundamental point is that expecting a significant part of the NTP user base to read the code simply does not scale, for a host of reasons. I will tell you that this is a combination of bits so it's not just a number. Each bit represents a test code that failed so you have quite a bit to look at. Just for curiosity, how many semicolons are there in the NTPv3 and NTPv4 codebases? My impression is that each is about 20,000 or 30,000, but I don't know why or where I got the number. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article rzidna3pv6cenclunz2dnuvz_o8la...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article 49bdbbbe.4030...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Joseph Gwinn wrote: What's the story for IBM's AIX? It builds on AIX too. It builds on most Unix systems though maybe not on some of the oldest O/S versions. Including the AIX we use, as mentioned in another posting. Joe Gwinn PS: I'll be offline for the next two weeks, starting tomorrow morning. If every post here results in an automated I'm out of the office message from you, we will never speak to you again! I did think of that ... but thought better of it. Nor did I trust Notes not to get into a fight with the reflector. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49bc631c.1060...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Ronan Flood wrote: On Thu, 12 Mar 2009 23:31:11 -0500, Joseph Gwinn joegw...@comcast.net wrote: NTP version 3 is running. I've been trying to find the command to give me the full version, including dot (like 3.4y), and I get answers, but don't know which one to believe, and if the version given is that of the NTP daemon itself, or of ntpq, or of ntpdate. It might get logged to /var/adm/messages or somewhere when xntpd starts, but try ntpq -c rv 0 daemon_version However we are not supporting ntp version 3, at least not without funding. Is there some reason why you are not running the latest version of ntpd? It's what came with that old version of Solaris, the most modern Solaris that will run on the old Sun boxes in question. Is NTP v4 proven to run on Solaris 9 (SunOS 5.9 Generic May 2002)? The suspicion is that we have not set something up correctly, not that NTP v3 has failed, or that NTP v4 would fare better or worse. Don't understand the comment about funding. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article lbgdns8bwol7mydunz2dnuvz_ukwn...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article 49bc631c.1060...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Ronan Flood wrote: On Thu, 12 Mar 2009 23:31:11 -0500, Joseph Gwinn joegw...@comcast.net wrote: NTP version 3 is running. I've been trying to find the command to give me the full version, including dot (like 3.4y), and I get answers, but don't know which one to believe, and if the version given is that of the NTP daemon itself, or of ntpq, or of ntpdate. It might get logged to /var/adm/messages or somewhere when xntpd starts, but try ntpq -c rv 0 daemon_version However we are not supporting ntp version 3, at least not without funding. Is there some reason why you are not running the latest version of ntpd? It's what came with that old version of Solaris, the most modern Solaris that will run on the old Sun boxes in question. Is NTP v4 proven to run on Solaris 9 (SunOS 5.9 Generic May 2002)? Proven? Please define that! It works for me. YMMV! OK, demonstrated. As opposed to it should work (but nobody has actually done it). So, you have done it, which is encouraging. The suspicion is that we have not set something up correctly, not that NTP v3 has failed, or that NTP v4 would fare better or worse. Don't understand the comment about funding. Simple enough. The people who maintain NTPD have to earn a living somehow. The more unpaid time they volunteer the more difficult it becomes to pay the bills! Open Source is doomed? Bill's prayers are answered at last. Version 3 is AT LEAST six years old. So is the OS version being used: 2009-2002= 7 years. So, they're siblings. V4.something is current. Vendors are still shipping V3 because the RFC for V4 has not yet been formally adopted. Or, if it has been adopted, the adoption is extremely recent. You forgot to mention laziness and pernicious inertia. If I recall, on a prior project, the version of NTP that came with the Solaris boxes of the day wasn't quite good enough (don't recall why), so the software folk downloaded and installed the then latest version of NTP (v3 I think) and that worked. The project team is still around. I think I'll chase the details down. If you have a C compiler, you can download the source and build it yourself. If you don't have a C compiler you can download GCC for free. Way too much work. It's never just compile and go. Significant futzing always seems necessary. Nor should this be necessary for Sun anything. But nagging at me is a half memory that on that prior project they may have had to compile NTP, for some possibly irrelevant reason. Like wanting to use the same toolchain for all code. Another reason to chase the details down. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49bd3907.1080...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Joseph Gwinn wrote: In article 49bc631c.1060...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: Ronan Flood wrote: On Thu, 12 Mar 2009 23:31:11 -0500, Joseph Gwinn joegw...@comcast.net wrote: NTP version 3 is running. I've been trying to find the command to give me the full version, including dot (like 3.4y), and I get answers, but don't know which one to believe, and if the version given is that of the NTP daemon itself, or of ntpq, or of ntpdate. It might get logged to /var/adm/messages or somewhere when xntpd starts, but try ntpq -c rv 0 daemon_version However we are not supporting ntp version 3, at least not without funding. Is there some reason why you are not running the latest version of ntpd? It's what came with that old version of Solaris, the most modern Solaris that will run on the old Sun boxes in question. Is NTP v4 proven to run on Solaris 9 (SunOS 5.9 Generic May 2002)? The suspicion is that we have not set something up correctly, not that NTP v3 has failed, or that NTP v4 would fare better or worse. Don't understand the comment about funding. Let me try and clarify my remark. NTP v3 was last released about 10 years ago. Since then all work and experience has been done on V4 at least in these forums. Sun has continued to ship V3 even though it's rather obsolete and old. One person within Sun is trying to change that but in the meantime users like you are left trying to deal with issues with that version. Very few of us in the forum have experience with V3 and can accurately answer questions about it. All of us have knowledge of V4 and what's true of V4 may very well be false for V3. Since everyone is a volunteer in this forum, unless they have a lot of spare time on their hands noone is going to look at the V3 sources and provide you with correct reponses for that version. Sun has support people who's job it is to answer such questions, but you of course pay for Sun support. If we had funding to do it, we would able to provide answers that you could rely on. Otherwise we are just guessing that things didn't change between V3 and V4 and there were fundamental changes between the two major versions. That was the reason for my remark. Hmm. OK, but I think that we've kind of run off the rails. Let me summarize: 1. Sun Microsystems' current behavior is not the issue, as I'm loading old software from an old CD onto old computer hardware, hardware that cannot support a newer version of Solaris than v9. One of these old Solaris boxes did work with NTPv3 running an even older version of Solaris, with no 5914 codes, deepening the mystery. The fact that this obsolete system can most likely support NTPv4 is worth investigation, though. 2. I think that what's happening is that I'm doing something dumb, and I bet that there is no real difference in how NTPv3 or NTPv4 would react to this faux pas, whatever it turns out to be. Nor is source code research needed or requested. 3. The original question was how to interpret a specific status code, 9514. I read the explanation in the documentation, but became no wiser for it. Thus my question. If there isn't a NTP FAQ entry on this, there probably should be. Our sysadmins were flummoxed by the cloud of 5914 codes, and they are far too busy to undertake a research project. (The deeper problem is that some managers believe that NTP is plug and play, which isn't quite true.) The various answers and questions I've gotten have been quite useful, as they give me a list of things to think about and investigate, things I might not have thought of, or soon thought of. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article ywn9r60yinft@ntp1.isc.org, Harlan Stenn st...@ntp.org wrote: In article joegwinn-795fe1.14394915032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net writes: Joseph In article 49bd3a1e.2020...@ntp.org, ma...@ntp.org (Danny Mayer) Joseph wrote: Danny NTP v4 builds on most versions of Solaris. If for some reason it does Danny not build, Harlan can help with that since he's the build master. We Danny have Solaris boxes in the build farm. I have access to a number of older versions of solaris for build testing. I have no idea now many people are running on these older OSes, but if there was a problem I believe I would have heard about it. I imagine so. Joseph What's the story for IBM's AIX? I have access to some AIX boxes for build testing. I generally do not have access to these boxes for runtime testing. No runtime testing? That's not good. But you'll no doubt hear when there's trouble. What AIX version and Technology Level (~=patch level) have been used to build NTPv4? One of the goals of the NTP Forum is to produce a diverse operational ntp build and test farm. This goal will be realized if a sufficient number of organizations join the NTP Forum. Version 3 is AT LEAST six years old. V4.something is current. Vendors are still shipping V3 because the RFC for V4 has not yet been formally adopted. Or, if it has been adopted, the adoption is extremely recent. NTPv3 was an RFC, but never a Standard. NTPv4 is, I believe, close to becoming a Standard. Joseph I have read at least one draft. The fact that it isn't yet a RFC is Joseph not a problem. Good. If you have a C compiler, you can download the source and build it yourself. If you don't have a C compiler you can download GCC for free. We use gcc for our Solaris builds. Joseph Also good to know, so I'll know better than to use the Sun compiler Joseph (not that it's bad, but that it isn't what's been worked through Joseph with NTP). I have platforms where I only use native compilers and not gcc. Again, these are build platforms, not operational test platforms. Yes. The context of the question was if I had to compile NTP from source, for whatever reason. Taming another toolchain is not something I would look forward to. And again, if there was a problem I believe I would have heard about it. Oh yes. The bullseye is invisible, but always there. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article nbwdnvxq_p-y_idunz2dnuvz_thin...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article 49bd3907.1080...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: [snip] 3. The original question was how to interpret a specific status code, 9514. I read the explanation in the documentation, but became no wiser for it. Thus my question. If there isn't a NTP FAQ entry on this, there probably should be. Our sysadmins were flummoxed by the cloud of 5914 codes, and they are far too busy to undertake a research project. (The deeper problem is that some managers believe that NTP is plug and play, which isn't quite true.) The various answers and questions I've gotten have been quite useful, as they give me a list of things to think about and investigate, things I might not have thought of, or soon thought of. Joe Gwinn Joe, You need to proofread your message text a little more carefully!! Which error are you ACTUALLY getting? You say 9514 and then 5914! Which is it? You're right, but it wouldn't help, for an odd reason. The status code is 9514. But I have a Clausing 5914 lathe. Inherent dyslexia inducer. Also, you might try Google with the FULL and EXACT text of the error message! It's 9514, pulled from a field in peerstats records. Think I'll get many false hits? Qualifying 9514 with peerstats brought me back to this thread. So, tried Maximum Distance Exceded, got led back to this exact news thread. But let's say I did find some relevant hits. This is the Internet. How would I know which hits to believe? The FAQ has to be the place for such explanations. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
Status codes also fixed below. In article ywn9mybmil8b@ntp1.isc.org, Harlan Stenn st...@ntp.org wrote: In article joegwinn-a341a8.15084015032...@news.giganews.com, Joseph Gwinn joegw...@comcast.net writes: Joseph Let me summarize: Joseph 1. Sun Microsystems' current behavior is not the issue, as I'm Joseph loading old software from an old CD onto old computer hardware, Joseph hardware that cannot support a newer version of Solaris than v9. Joseph One of these old Solaris boxes did work with NTPv3 running an even Joseph older version of Solaris, with no 9514 codes, deepening the mystery. Joseph The fact that this obsolete system can most likely support NTPv4 is Joseph worth investigation, though. I believe ntp4 will work there, and if it does not and somebody opens a bug report on it, I expect it will be fixed. Yes. Joseph 2. I think that what's happening is that I'm doing something dumb, Joseph and I bet that there is no real difference in how NTPv3 or NTPv4 Joseph would react to this faux pas, whatever it turns out to be. Nor is Joseph source code research needed or requested. I mostly agree with you here, more in a bit... Joseph 3. The original question was how to interpret a specific status Joseph code, 9514. I read the explanation in the documentation, but became Joseph no wiser for it. Thus my question. Joseph If there isn't a NTP FAQ entry on this, there probably should be. Joseph Our sysadmins were flummoxed by the cloud of 9514 codes, and they Joseph are far too busy to undertake a research project. (The deeper Joseph problem is that some managers believe that NTP is plug and play, Joseph which isn't quite true.) I think you are talking about one of my pet peeves: http://support.ntp.org/bin/view/Dev/NtpVariablesAndNtpq I don't think that I have inconsistent versions of ntpd and ntpq, because both came off the same CD from Sun Microsystems. I strongly believe that we should implement something like this. It will need to be implemented in a way that Dave can tolerate. The odds of this being implemented are directly proportional to somebody doing the work. Near as I can tell, the best way to have somebody do the work is to have money available to pay for the work to be done. The best way I know to get money to pay for this work to be done is to show organizations that by joining the NTP Forum they will be spending money to get significant value in return. I believe projects like this one are one example of that significant value. It would be nice for sure, but I cannot see companies not selling time-related equipment having a sufficient business case to fund the NTP Forum. Being a user (versus maker) of such equipment is not generally sufficient. This analysis is not restricted to time-related stuff. I spent many years working on POSIX standards. The only reason my employer was willing to support this effort (and the associated travel expenses) was that our customers demanded conformance to such standards, and also that we help develop those standards. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article qo6dnzljdv8z4cdunz2dnuvz_hwwn...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article nbwdnvxq_p-y_idunz2dnuvz_thin...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: In article 49bd3907.1080...@ntp.org, ma...@ntp.org (Danny Mayer) wrote: [snip] 3. The original question was how to interpret a specific status code, 9514. I read the explanation in the documentation, but became no wiser for it. Thus my question. If there isn't a NTP FAQ entry on this, there probably should be. Our sysadmins were flummoxed by the cloud of 5914 codes, and they are far too busy to undertake a research project. (The deeper problem is that some managers believe that NTP is plug and play, which isn't quite true.) The various answers and questions I've gotten have been quite useful, as they give me a list of things to think about and investigate, things I might not have thought of, or soon thought of. Joe Gwinn Joe, You need to proofread your message text a little more carefully!! Which error are you ACTUALLY getting? You say 9514 and then 5914! Which is it? You're right, but it wouldn't help, for an odd reason. The status code is 9514. But I have a Clausing 5914 lathe. Inherent dyslexia inducer. Also, you might try Google with the FULL and EXACT text of the error message! It's 9514, pulled from a field in peerstats records. Think I'll get many false hits? Qualifying 9514 with peerstats brought me back to this thread. So, tried Maximum Distance Exceded, got led back to this exact news thread. But let's say I did find some relevant hits. This is the Internet. How would I know which hits to believe? I would be influenced by who wrote it and who disagreed with him! But what if I listen to the loudest one? The FAQ has to be the place for such explanations. I'm not sure if this qualifies as an FAQ as I don't recall that it has come up before. FAQ stands for Frequently Asked Questions. RAQ then? Rarely Asked Questions Seriously, I can't believe that I'm the only person in history to be perplexed by these status codes, and those little three-word summaries are a bit telegraphic. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49bb8860$0$507$5a6ae...@news.aaisp.net.uk, David Woolley da...@ex.djwhome.demon.co.uk.invalid wrote: Joseph Gwinn wrote: In article 49bae109$0$505$5a6ae...@news.aaisp.net.uk, Da ntpd doesn't care about what the drift is in determining root distance. It simply takes the position that the actual local clock will be somewhere within +/- 15ppm of the value which would achieve perfect phase lock with true time. This part seems to conflict with the max +/- 500 ppm steering authority of NTP. How are these two limits related? They are not. The 500ppm is the range of correction that can be applied. The 15ppm is a pessimistic estimate of the error in setting that correction. I.E. if there is a valid time source, ntpd may decide it needs a correction of 300ppm. In the absence of that time source, it assumes that the correction it really needed was between 285ppm and 315ppm, with the uncertainty being due to measurement error and changes in the local clock frequency. That uncertainty in frequency causes and uncertainty in time which grows with time, until it, when combined with other uncertainties, exceeds 1 second. The client compares the uncertainty for each server with one second, and when that is exceeded starts ignoring the server. I understand this, but what perplexes me is that both timeservers, of different make and model, are showing the same behavior, so I have to believe that the client is somehow not right. My theory is that we have not succeeded in getting a clean and correct startup. By the way, the datasets span at least 24 hours, no startup transient is seen, and the behavior does not appear to change over the run. So there may also be a configuration error, but it's hard to imagine what could do this. The assumed maximum reasonable error therefore grows at 15 microseconds per second. One suspicion I have is that the drifts file has data from some other test still in it. We will try deleting the drifts file. Root distance exceeded is a problem with the server, not the client. Usually, but the fact that both servers are rejected makes me wonder, as discussed above. Another suspicion is that the computer's sense of time is too far away from that provided by the timeserver. I would have thought this would cause the daemon to balk and complain, but perhaps there is a window where it will not balk but will struggle mightily. We will use ntpdate as part of the startup process and see if it matters. This will cause ntpd to terminate. I repeat, it is your servers that are being rejected. Yes. The question is why. For the record: In this test setup, at any given time, each client (daemon) sees at most one timeserver. All GPS receivers are fed from a single roof antenna by a splitter. The mapping between client and server changes only between test runs. Each run lasts at least 24 hours, and may run over a weekend. Generally, in this sort of case, it is helpful to have output from ntpq's peers, assoc and rv commands, with the latter for each association number from assoc and for 0, i.e. the machine itself. That should tell us exactly when the servers had valid time, etc. Yes. I'll collect this data on Monday. I have been suspicious of the old Symmetricom ET6010 GPS receiver in the lab before. Elsewhere in the large system we have observed that sometimes one must reset the ET6010 to get good time fed to the TS2100 timeserver; no idea why. But the fact that the brand new Spectracom 9383 timeserver cum GPS receiver does the same thing caused focus to shift elsewhere. By the way, all GPS receivers discussed have Rubidium local oscillators. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
[ntp:questions] What exactly does Maximum Distance Exceded mean?
I have been debugging some system problems. The main system is too complicated, with too many people doing too many things, so I sought quiet refuge in an isolated test system consisting of a NTP timeserver connected by a point-to-point ethernet cable to a computer running NTP, which generates peerstats and loopstats data. This test system is air-gap isolated from the rest of everything. Only one timeserver is available to a given computer at a time. The timeserver can be either a Symmetricom ET6010 GPS receiver feeding an IRIG-B002 time signal to a Symmetricom TS2100 Network Time Server, or a Spectracom 9383 NTP timeserver with built-in GPS receiver. The GPS receivers are driven from a common antenna via a splitter. The computer can be a Sun Ultra 10 or a Sun Ultra 60, in both cases running Solaris 9. Solid boxes, but old. The OS version reply is SunOS 5.9 Generic May 2002. This was clean installed from CD a week ago, so has not had time to collect too many barnicles. NTP version 3 is running. I've been trying to find the command to give me the full version, including dot (like 3.4y), and I get answers, but don't know which one to believe, and if the version given is that of the NTP daemon itself, or of ntpq, or of ntpdate. The full grid of four tests, being two timeservers by two computers, has been run. Many odd things are seen, but the question for today is about status codes in peerstats file records. Most of the replies that NTP is using to update the time have a status code of 9514, which translates to the following: Configured, reachability OK; Current sync source - max distance exceeded; Count is 1; Peer now reachable. The part that has me most perplexed is the max distance exceeded part, as this is a direct wired connection, with zero hops, zero delay, and no interfering traffic. Obviously, they are not talking about physical distance or hops or the like, so the distance has to have units of time. Although most received replies have status 9514, they are nonetheless used to update the loop filter and so appear in the loopstats file. When I co-plot loopstats and peerstats, the loopstats dots land on top of the peerstats dots. What is this error likely telling me? What are the possibilities? What tests will tell the tale? Thanks to all, Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 8dca286d-ba66-4caf-8090-a3467064a...@s20g2000yqh.googlegroups.com, Mike K Smith mks-use...@dsl.pipex.com wrote: On 13 Mar, 04:31, Joseph Gwinn joegw...@comcast.net wrote: The timeserver can be either a Symmetricom ET6010 GPS receiver feeding an IRIG-B002 time signal to a Symmetricom TS2100 Network Time Server, or a Spectracom 9383 NTP timeserver with built-in GPS receiver. The GPS receivers are driven from a common antenna via a splitter. I'm familiar with the Spectracom 9383, but not the other equipment. The Symmetricom units are quite old. What does the NTP status page show? What does the GPS Signal Status show? How many satellites are you seeing? Is the device reporting 'Position Hold'? I'll check. Most of the replies that NTP is using to update the time have a status code of 9514, which translates to the following: Configured, reachability OK; Current sync source - max distance exceeded; Count is 1; Peer now reachable. The part that has me most perplexed is the max distance exceeded part, as this is a direct wired connection, with zero hops, zero delay, and no interfering traffic. Obviously, they are not talking about physical distance or hops or the like, so the distance has to have units of time. Although most received replies have status 9514, they are nonetheless used to update the loop filter and so appear in the loopstats file. You say most are 9514, are there any 96xx values? The status codes seen are 9014 (red), 9514 (orange), and 9614 (green). The colors are those of the plotted dots. Loopstats dots are dark blue and smaller, so when co-plotted one sees little bullseyes. I see 9014 when changing cables or timeservers, and in most other tests I see mostly 9614 and a few 9014, and very rarely 9514 until now. When I co-plot loopstats and peerstats, the loopstats dots land on top of the peerstats dots. I've never seen a device come up as status 5, but since the RFC1305 text treats it as current synchronization source; max distance exceeded (if limit check implemented) then I guess it makes sense that it will use it as the sync source and will update loopstats appropriately. I had not seen 5 before either, but NTP is clearly using these replies to update the time. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article 49ba0e33$0$505$5a6ae...@news.aaisp.net.uk, David Woolley da...@ex.djwhome.demon.co.uk.invalid wrote: Joseph Gwinn wrote: What is this error likely telling me? What are the possibilities? What tests will tell the tale? Your timeservers are unsynchronised, but for some reason not setting their stratum to 16. Distance exceeded means that the combination of worst case round trip time induced error and an assumed drift of 15ppm since the last valid time on the root server (plus a few minor components) has exceeded 1 second. How would it know of drift, in this isolated little island? Perhaps the drift file is causing trouble. Perhaps ntpdate was needed to get things started properly started. I bet the engineer did not do any of these things. He just plugged thing together. Some of the more recent loopstats and peerstats files have data from both timeservers, the transition being marked by a little cloud of 5014 status (red dots in my plots). It commonly happens with w32time servers that have been synchronized once but left to drift. It can also happen if the servers are orphan mode, and haven't had a real time source for too long, and you are not using the very latest orphan mode code. This strengthens my impression that we don't have a clean startup, so clean startup will be the next thing to try. We were equally sloppy before, but didn't get the cloud of 9514 complaints, probably because until now things weren't quiet enough for NTP to worry about the change. Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What exactly does Maximum Distance Exceded mean?
In article enydnx61waj9gsbunz2dnuvz_tjin...@giganews.com, Richard B. Gilbert rgilber...@comcast.net wrote: Joseph Gwinn wrote: I have been debugging some system problems. The main system is too complicated, with too many people doing too many things, so I sought quiet refuge in an isolated test system consisting of a NTP timeserver connected by a point-to-point ethernet cable to a computer running NTP, which generates peerstats and loopstats data. This test system is air-gap isolated from the rest of everything. Only one timeserver is available to a given computer at a time. The timeserver can be either a Symmetricom ET6010 GPS receiver feeding an IRIG-B002 time signal to a Symmetricom TS2100 Network Time Server, or a Spectracom 9383 NTP timeserver with built-in GPS receiver. The GPS receivers are driven from a common antenna via a splitter. The computer can be a Sun Ultra 10 or a Sun Ultra 60, in both cases running Solaris 9. Solid boxes, but old. The OS version reply is SunOS 5.9 Generic May 2002. This was clean installed from CD a week ago, so has not had time to collect too many barnicles. NTP version 3 is running. I've been trying to find the command to give me the full version, including dot (like 3.4y), and I get answers, but don't know which one to believe, and if the version given is that of the NTP daemon itself, or of ntpq, or of ntpdate. The full grid of four tests, being two timeservers by two computers, has been run. Many odd things are seen, but the question for today is about status codes in peerstats file records. Most of the replies that NTP is using to update the time have a status code of 9514, which translates to the following: Configured, reachability OK; Current sync source - max distance exceeded; Count is 1; Peer now reachable. The part that has me most perplexed is the max distance exceeded part, as this is a direct wired connection, with zero hops, zero delay, and no interfering traffic. Obviously, they are not talking about physical distance or hops or the like, so the distance has to have units of time. I think that, perhaps, maximum distance refers to synchronization distance q.v. Once upon a time, I knew the definition but my memory has failed me. The other answers didn't use this exact term, but it sounds like the same idea. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] NTP over redundant peer links, undetected loops
In article 5d7f07420902151105m48a5e210s72e8e168e67d1...@mail.gmail.com, malay...@gmail.com (Ryan Malayter) wrote: On Sun, Feb 15, 2009 at 12:23 PM, Danny Mayer ma...@ntp.org wrote: Because I want to get away from the notion that these are meant to be IP addresses. In addition in an IPv6-only environment that wouldn't work either. Why create work when it's unnecessary just to find a valid IP address? In addition with anycast addresses are not globally unique. The chances that you will create a non-unique random number within a network is extremely low. It depends on the size of the network. The chances of a duplicate 32-bit number on a network including 65000 hosts is about 40%. The NTP Pool network, which comprises at least 10^6 hosts, for example, would have collision probability very close to 1. How did you compute that? Given that 2^32= ~4*10^9, it's hard to see how 10^6 hosts spread at random in a 10^9 codespace could achieve 100% collision probability. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Uwe Klein [EMAIL PROTECTED] wrote: Hal Murray wrote: I did get a look at the ntpd script today. Turns out the answer on where it gets the ntp.conf file is right there, near the top, in the line ntpconf=/etc/ntp.conf, even though the ntp man page points us deeper in the /etc hierarchy. The sysadmin I was working with was real annoyed, as the misinformation in the man page had sent him into circles. We will add pointer comments to all placebo ntp.conf files, to save future generations of sysadmins from this fate. I still don't know which ntp.conf you are really using. I'm looking at a Fedora 6 box. If you look in /etc/init.d/ntpd, you will see that it mucks about with ntpconf (the one above) to find the servers. Those servers get passed to ntpdate. Mumble. That's old crap. There is now a command line switch that does the right thing. I don't see where ntpconf gets passed to ntpd as a command line argument. If the man page says ntpd uses some other config file, it is probably right, or at it seems to me that it would be more likely that the guy who changed the code also changed the man page but didn't fixup the init script. Does Red Hat write distribution specific manpages? I would be surprised if. Well, there was a full man page for ntp on RHEL, one that's far longer than your example below, and someone wrote it. Don't know who, but the RHEL box was bought from IBM, who are famous for their documentation, so I would venture that IBM augmented the man page over what Red Hat provides. And the E in RHEL is Enterprise, and enterprises want full documentation delivered with the product. Joe Gwinn uwe This is the recent SuSE Linux Manpage that all ntp related keywords point to: NTP(1) NTP(1) NAME NTP - Network Time Protocol SEE ALSO The NTP distribution does not include man pages. To learn more about the NTP protocol and this software, please install the xntp-doc package included in you SuSE Linux dis- tribution. In /usr/share/doc/packages/xntp-doc you will find the complete set of documentation on building and configuring a NTP server or client. The documentation is in the form of HTML files suitable for browsing and contains links to additional documentation at var- ious web sites. Also included: What about NTP? Understanding and using the Network Time Protocol. A first try on a non-technical Mini-HOWTO and FAQ on NTP. Edited by Ulrich Windl and David Dalton. Further information on NTP in the Internet can be found in the NTP web page at http://www.eecis.udel.edu/~ntp/ ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Hal Murray) wrote: I did get a look at the ntpd script today. Turns out the answer on where it gets the ntp.conf file is right there, near the top, in the line ntpconf=/etc/ntp.conf, even though the ntp man page points us deeper in the /etc hierarchy. The sysadmin I was working with was real annoyed, as the misinformation in the man page had sent him into circles. We will add pointer comments to all placebo ntp.conf files, to save future generations of sysadmins from this fate. I still don't know which ntp.conf you are really using. I'm looking at a Fedora 6 box. If you look in /etc/init.d/ntpd, you will see that it mucks about with ntpconf (the one above) to find the servers. Those servers get passed to ntpdate. Mumble. That's old crap. There is now a command line switch that does the right thing. I don't see where ntpconf gets passed to ntpd as a command line argument. Next time I have a hands-on session, I'll put something in the purported correct ntp.conf file to tell if this file is in fact the Chosen One. If the man page says ntpd uses some other config file, it is probably right, or at it seems to me that it would be more likely that the guy who changed the code also changed the man page but didn't fixup the init script. RHEL via IBM and Fedora may or may not be identical, even though both ultimately came from Red Hat. As discussed in another posting, a full NTP man page comes with RHEL via IBM, unlike Fedora. The details of how NTP is managed and run may also have been improved. The intent of the service utility is to simplify the day-today activities of the sysadmins, and it probably succeeds. The root problem is turning out to be with the documentation of service and ntp, not with service and ntp themselves. It is pretty clear that the RHEL via IBM man page for NTP is incorrect. This would not be the first time in history that documentation got out of step with code. I'll smoke this out by filing a formal bug report with IBM against the ntp man page. This cost us some real money, not just annoyance, and we cannot be alone. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Bill Unruh [EMAIL PROTECTED] wrote: Joseph Gwinn [EMAIL PROTECTED] writes: In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Hal Murray) wrote: I'm not a sysadmin, but am digging into service. I don't recall that the service man page was that helpful, but will look again. service is mostly a shortcut to save typing. If you think it is getting in your way, run /etc/init.d/ntpd whatever by hand. (It also fixes up environment and cd-ed directory and whatever.) Yes. This is what we did to prove that NTP really could generate loopstats and peerstats. No I suspect you ran /usr/sbin/ntpd, not /etc/init.d/ntpd /etc/init.d/ntpd start should do EXACTLY the same thing as when the system runs it on bootup. If I recall, the line that worked was /etc/init.d/ntpd -c filename of our ntp.conf file. I don't recall that sbin was involved. The -x command to bash will print each line as it gets expanded and executed. So you might try something like: bash -x /etc/init.d/ntpd start to see what is really going on. Another good idea to try. It of course produces far more output but obviates the need to insert echo lines into /etc/init.d/ntpd Yep. But if it solves the problem, I won't mind the blather. Note, I am wondering what has happened to all these suggestions? Have you tried any of them yet? Have you discovered what it is actually using as its configuration file? I have been collecting all the suggestions I have heard here, and will try them when the relevant sysadmin is able to spare the time. This may be today (Friday). Unless he is somehow deflected. You might want to post the config file here (ntp.conf) here in case it is some error in that file which is causing your problems rather than that ntpd is using some other config file. We are happily collecting loopstats and peerstats data on RHEL using that ntp.conf file, once we started the daemon manually with an explicit filepath argument (as described above), so the ntp.conf file itself does not appear to be the problem. I was suspicious of that file too, and so had cleaned it down to something like three lines, basically following the minimum ntp.conf example given in the online NTP documentation. By the way, I doubt that it matters here, but this RHEL is running NTPv4. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] What happens if ntp server unavailable at start up?
In article [EMAIL PROTECTED], Unruh [EMAIL PROTECTED] wrote: Harlan Stenn [EMAIL PROTECTED] writes: [snip] Unruh Did the dynamic keyword ever work? The web docs say that it is not Unruh yet implimented. I'm pretty sure it works - what documentation says it doesn't? Some document on ntp.org describing the options in the ntp.conf file. I do not want to look for it again-- the docs are incredibly hard to search-- one of the problems with making them into an infinite number of web pages. What I find useful is to use google with a site:udel.edu qualifier. It would help if the docs entire tree were under ntp.org though. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Hal Murray) wrote: I'm not a sysadmin, but am digging into service. I don't recall that the service man page was that helpful, but will look again. service is mostly a shortcut to save typing. If you think it is getting in your way, run /etc/init.d/ntpd whatever by hand. (It also fixes up environment and cd-ed directory and whatever.) Yes. This is what we did to prove that NTP really could generate loopstats and peerstats. The -x command to bash will print each line as it gets expanded and executed. So you might try something like: bash -x /etc/init.d/ntpd start to see what is really going on. Another good idea to try. Thanks, Joe ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Unruh [EMAIL PROTECTED] wrote: David Woolley [EMAIL PROTECTED] writes: James Cloos wrote: I read through most of the replies so far, but one thing I haven't seen noted is that this isn't an ntp issue at al Did you mean service (8). Treating it as a black box is basically how Red Hat is marketed; it is basically in the same market as Windows. People who want a white box Linux are more likely to choose something like Slackware. Agree. service is a dead simple program. It runs its argument from the /etc/init.d directory. Anyway, long ago we suggested that he looked in /etc/init.d/ntpd to see if there was anything in there that suggested which config file was being used. I found the script, and started reading it, and will return to it. Or insert an echo where ... is the exact line that that script runs to start ntpd to see if there are any interesting arguments to ntpd. Or put in an env in there to see exactly what the environment is that ntpd sees. All good ideas. Direct, and free of excess assumptions that things are as things should be. It is just a damn shell script. It is not a black box. I've been reading it, and it does seem simple, but haven't really studied it yet. I think the OP has gone far beyond what the average RHEL administrator is expected to do in terms of looking inside the box. That's for sure, and is why I'm doing the debugging, even though I'm not a sysadmin. The sysadmins really don't understand NTP. I find it extremely unlikely that RHEL uses anything but /etc/ntpd.conf but if it does then it is up to Redhat to document it. I suspect either user error or some admin in the past of this organization has changed things and never documented it. Judging by the cruft accumulation in the trojan ntp.conf file, this is not a virgin install, so I'd bet on confused sysadmins. They may have been trying to get it to work, not realizing that this ntp.conf file is only a placebo. We have not had a report back from him as to what the results were of all the suggestions we made. Because there is nothing to report yet, due to the press of other business. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Unruh [EMAIL PROTECTED] wrote: Joseph Gwinn [EMAIL PROTECTED] writes: In article [EMAIL PROTECTED], Steve Kostecke [EMAIL PROTECTED] wrote: On 2008-09-03, Joseph Gwinn [EMAIL PROTECTED] wrote: Read the service shell script. It appears to get its file paths from environment variables named after the thing being started and stopped and accessible only in the root environment; this bit of RHEL-specific structure is being chased down. (Does anyone know where this is documented?) On Linux OSes init scripts are typically found in /etc/init.d/ or /etc/rc.d/init.d/ Look for one named ntp (or something containing ntp). Yes, and that's where strace led me, where I found a script called ntpd. How the service script interacts with this ntpd script isn't clear. Environment variables seem to be implicated, but a listing of environment variables is not helpful. Next week I'll digest it all. service simply runs the program listed as its argument from the /etc/init.d directory. Ie, service ntpd start is the same as /etc/init.d/ntpd start True, but there seems to be more to it than that. Next week. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Unruh [EMAIL PROTECTED] wrote: Joseph Gwinn [EMAIL PROTECTED] writes: In article [EMAIL PROTECTED], Peter J. Cherny [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: ... Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without ... [EMAIL PROTECTED] ~]$ strings /usr/sbin/ntpd|grep ntp.conf /etc/ntp.conf In the RHEL case, this would find exactly the wrong copy of ntp.conf, being the one we were changing to no avail, not the one that NTP was in fact using. Which one was ntp in fact using? Don't know yet. Other than it wasn't the obvious one. When we do figure it out, all pretenders to the throne will be summarily deleted, to prevent confusion. [EMAIL PROTECTED] ~]$ strace -f -o x /usr/sbin/ntpd -g I'll have to look into this. It sounds like it might be general enough. [EMAIL PROTECTED] ~]# grep ntp.conf x 3351 open(/etc/ntp.conf, O_RDONLY) = 4 Doesn't this assume that the correct ntp.conf file is called ntp.conf? It may be common, the standard convention, but it is not required. The whole point is to find the correct file without making assumptions, because on a strange computer strange things may have been done. yes, but then do strace as above and look through the file looking for something that might be a configuration file. If they call it /lib/libc.so then you are probably shit out of luck, but usually they will not do that. The strace gave a lot of data, mostly irrelevant, which I will plow through next week. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Steve Kostecke [EMAIL PROTECTED] wrote: On 2008-09-03, Joseph Gwinn [EMAIL PROTECTED] wrote: Read the service shell script. It appears to get its file paths from environment variables named after the thing being started and stopped and accessible only in the root environment; this bit of RHEL-specific structure is being chased down. (Does anyone know where this is documented?) On Linux OSes init scripts are typically found in /etc/init.d/ or /etc/rc.d/init.d/ Look for one named ntp (or something containing ntp). Yes, and that's where strace led me, where I found a script called ntpd. How the service script interacts with this ntpd script isn't clear. Environment variables seem to be implicated, but a listing of environment variables is not helpful. Next week I'll digest it all. Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without being able to find or see the actual command line or lines that launched the daemon? I did not see a ntpq command that sounded plausible, although ntpq would be an obvious choice. This would be very useful for debugging, as each and every platform type seems to have a different approach to handling NTP. Why not use the file location features built in to your OS to find all possible instances of ntp.conf? $ locate ntp.conf or $ find / -name ntp.conf Pipe the output of either of those commands to 'xargs ls -l' to see the datestamps of the files. We did this, but could not tell which one mattered. Next week. Nor is it *required* the the ntp configuration file be called ntp.config. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], David Woolley [EMAIL PROTECTED] wrote: Steve Kostecke wrote: On 2008-09-03, Joseph Gwinn [EMAIL PROTECTED] wrote: Read the service shell script. It appears to get its file paths from environment variables named after the thing being started and stopped and accessible only in the root environment; this bit of RHEL-specific structure is being chased down. (Does anyone know where this is documented?) On Linux OSes init scripts are typically found in /etc/init.d/ or /etc/rc.d/init.d/ Look for one named ntp (or something containing ntp). I believe service is just a front end to those scripts, so I presume that, by service shell scripts he is referring to those scripts. The problem he is having is that they probably source files (bash . command) files containing shell variable definitions from the master configuration directory, maintained by the, typically GUI, configuration tools. I suspect he hasn't realised that is is sourcing thesse files. Ahh. I had figured out the first part of this, but had not figured out where the data was kept. Environment variables didn't have anything plausible. But it has to come from *somewhere*. The sysadmins know nothing of all this, being AIX and Solaris guys. Note that not all Linux distributions use this style of startup script, some are based on a more historical style of /etc/rc. Natch. That's why ntpq needs a bit more built-in debug support. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Richard B. Gilbert [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: In article [EMAIL PROTECTED], Martin Burnicki [EMAIL PROTECTED] wrote: Joe, Joseph Gwinn wrote: In article [EMAIL PROTECTED], Peter J. Cherny [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: ... Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without ... [EMAIL PROTECTED] ~]$ strings /usr/sbin/ntpd|grep ntp.conf /etc/ntp.conf In the RHEL case, this would find exactly the wrong copy of ntp.conf, being the one we were changing to no avail, not the one that NTP was in fact using. [EMAIL PROTECTED] ~]$ strace -f -o x /usr/sbin/ntpd -g I'll have to look into this. It sounds like it might be general enough. [EMAIL PROTECTED] ~]# grep ntp.conf x 3351 open(/etc/ntp.conf, O_RDONLY) = 4 Doesn't this assume that the correct ntp.conf file is called ntp.conf? It may be common, the standard convention, but it is not required. The whole point is to find the correct file without making assumptions, because on a strange computer strange things may have been done. I fully agree. Ntpd generates a bunch of messages about what it has found in the config file, at least in debug mode. Maybe you should open an enhancement request on http://bugs.ntp.org to make ntpd also print the name of the config file it is using, maybe only in debug mode. I'm surprised that it doesn't already print the full filename of every file it uses. Will debug mode do much if the binary wasn't compiled for debug? I'm trying to use the provided binary, whatever it might be, and recompiling is usually far too much trouble to be practical. Especially as the effort is per platform type, and we have multiple types. I will file an enhancement request. However, my feeling is that this function would be most useful if added to ntpq, and yielded the full filename including directories, as there may be multiple ntp.conf files scattered about. The key is to get NTP to tell us which file NTP is using, without interference from our firmly held but sadly mistaken assumptions about what NTP ought to be doing. Joe Gwinn Since the source to NTPD is available, it's a SMOP to modify it to print out the desired file specification! True enough, but far too much work. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Martin Burnicki [EMAIL PROTECTED] wrote: Richard B. Gilbert wrote: ISTR that ntpd looks in /etc/inet if it is not told to look elsewhere by the command that starts ntpd. This should take care of Unix and Unix-like systems. Windoze?? Ask someone who knows. AFAIK this is the default location under Solaris, but e.g. under Linux the location is just /etc. Anyway, this is configured at compile time and maybe overridden by a command line parameter, in which case it does not help to know the default. On some systems the command line parameters are displayed in the process list, so you can: 1.) Look at the process list to see if a configuration file has been specified 2.) If it has not, grep through the ntp binary to find the path of the default config file 3.) see if that file exists Please note that especially under Windows things may look different. The NTP service first tries to open %windir%\ntp.conf, and, if that file does not exist, %windir%\system32\drivers\etc\ntp.conf. The GUI installers provided by Meinberg override these settings with an etc\ directory below the program installation path, by default \program files\ntp\etc. The configured setting can be retrieved from the ImagePath registry key of the NTP service registry entry. If you are upgrading an installation of NTP under Windows then there may still be old config files under the older paths, so you have to look explicitely which of the file has being read by the running NTP service. If ntpd would write a log message at startup then you could easily find out on every platform which config file has been read. That would certainly work, and work in all cases. My problem is to debug NTP problems in multiple systems that I have limited knowledge of, ones that may or may not follow the usual conventions, or the same conventions, and which may in fact may have been hosed up by some sysadmin who knows nothing of NTP save where the big red start button is supposed to be. To be useful in such an environment, debug tools must be platform independent and cannot make assumptions about conventions being followed. I am not worried about the case where someone compiles their own munged version of NTP. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Peter J. Cherny [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: I will file an enhancement request. However, my feeling is that this function would be most useful if added to ntpq, and yielded the full filename including directories, as there may be multiple ntp.conf files scattered about. The key is to get NTP to tell us which file NTP is using, without interference from our firmly held but sadly mistaken assumptions about what NTP ought to be doing. And you'll file enhancement requests for every other daemon on the machine ??? Only the ones that sufficiently annoy me. In most OSs the man pages are definitive and mostly correct, with changes noted in the release notes. If you've paid your support fees, ask the vendor. In most of the Unix family, the source is available. A support question is being placed with IBM. But newsgroups often think of angles that support does not. flame Else, ask a SysAdmin/SysEngineer/SwEngineer who does this for a living. You do have other than junior staff ? /flame Umm. I'm not a sysadmin. If we had such a person, I wouldn't be doing the debugging. Few people even know that NTP exists, let alone how it works well enough to debug an installation. When there is a time problem, the sysadmins come and grab me to help them. So far, I've always been able to figure the problem out. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file (Joseph Gwinn)
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Breck Beatie) wrote: This isn't quite what you're asking for and it's certainly not ntp specific, but one technique that I have used in the past is to replace the binary I'm trying to debug with a script which dumps useful information and then forwards the exec to the real binary. This should work, but is a bit of work. I'll keep it in reserve. I usually have it dump its environment and the full set of command line arguments someplace safe and then exec the original binary. You could certainly have it run the original binary with strace. I'm going to grind through the strace output next week. I have friends who'll run the binary with gdbserver and then they connect with gdb have their way with the binary. I've never done that so I have no idea how you'd invoke gdbserver. I don't know if we even have gdbserver. Joe Gwinn Joe Gwinn writes: Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without being able to find or see the actual command line or lines that launched the daemon? I did not see a ntpq command that sounded plausible, although ntpq would be an obvious choice. This would be very useful for debugging, as each and every platform type seems to have a different approach to handling NTP. ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Richard B. Gilbert [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: In article [EMAIL PROTECTED], Steve Kostecke [EMAIL PROTECTED] wrote: On 2008-09-03, Joseph Gwinn [EMAIL PROTECTED] wrote: Read the service shell script. It appears to get its file paths from environment variables named after the thing being started and stopped and accessible only in the root environment; this bit of RHEL-specific structure is being chased down. (Does anyone know where this is documented?) On Linux OSes init scripts are typically found in /etc/init.d/ or /etc/rc.d/init.d/ Look for one named ntp (or something containing ntp). Yes, and that's where strace led me, where I found a script called ntpd. How the service script interacts with this ntpd script isn't clear. Environment variables seem to be implicated, but a listing of environment variables is not helpful. Next week I'll digest it all. Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without being able to find or see the actual command line or lines that launched the daemon? I did not see a ntpq command that sounded plausible, although ntpq would be an obvious choice. This would be very useful for debugging, as each and every platform type seems to have a different approach to handling NTP. Why not use the file location features built in to your OS to find all possible instances of ntp.conf? $ locate ntp.conf or $ find / -name ntp.conf Pipe the output of either of those commands to 'xargs ls -l' to see the datestamps of the files. We did this, but could not tell which one mattered. Next week. Nor is it *required* the the ntp configuration file be called ntp.config. Joe Gwinn There MIGHT, in rare cases, be good reason NOT to call the configuration file ntp.conf (it's conf not config, unless someone changed it recently). IF so, both the new name and the reasons for it should be documented! In most cases it's best to stick with the de facto standard. I agree completely. But I didn't set the thing up. But I do have to figure it out and fix it. And document it. It did flummox all our sysadmins, although as with sysadmins worldwide they are too busy. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Peter J. Cherny [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: ... Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without ... [EMAIL PROTECTED] ~]$ strings /usr/sbin/ntpd|grep ntp.conf /etc/ntp.conf In the RHEL case, this would find exactly the wrong copy of ntp.conf, being the one we were changing to no avail, not the one that NTP was in fact using. [EMAIL PROTECTED] ~]$ strace -f -o x /usr/sbin/ntpd -g I'll have to look into this. It sounds like it might be general enough. [EMAIL PROTECTED] ~]# grep ntp.conf x 3351 open(/etc/ntp.conf, O_RDONLY) = 4 Doesn't this assume that the correct ntp.conf file is called ntp.conf? It may be common, the standard convention, but it is not required. The whole point is to find the correct file without making assumptions, because on a strange computer strange things may have been done. Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Finding out where ntpd gets its ntp.conf file
In article [EMAIL PROTECTED], Richard B. Gilbert [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: We had been struggling with NTP running on Red Hat Enterprise Linux (RHEL) on IBM-built Intel boxes, specifically with getting NTP to generate loopstats and peerstats files. Basically, nothing worked, despite many attempts. Yesterday, I cracked it while working on the problem with one of the sysadmins. The ntp.config file was very complex, and I suspected it of being mostly flotsam and jetsam from prior uses, and probably in conflict with itself, so we cleaned it down to maybe three lines, and then stopped and started ntpd using the service utility. The daemon started, but complained that it was unable to synchronize. The sysadmin mentioned that it very often did this, for unknown reasons. Then I noticed that the timeserver IP address was not the same as specified in the simplified ntp.conf file, and sure enough the address that NTP was trying to use was not accessible to ping. Hmm. Huh? NTP cannot be using the ntp.conf file we thought it was. Tried starting NTP manually with -c option and providing the full path to our ntp.conf file. Success! Read the service shell script. It appears to get its file paths from environment variables named after the thing being started and stopped and accessible only in the root environment; this bit of RHEL-specific structure is being chased down. (Does anyone know where this is documented?) Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without being able to find or see the actual command line or lines that launched the daemon? I did not see a ntpq command that sounded plausible, although ntpq would be an obvious choice. This would be very useful for debugging, as each and every platform type seems to have a different approach to handling NTP. Joe Gwinn I don't recall ever encountering such a facility. Or ever needing one. You are very fortunate. I do need one. You have confirmed my suspicion that NTP has no such facility. Use of strace has been suggested. It is on some but not all platforms at present. It seems to me that this is the sort of thing that the sysadmin should be documenting. And if he has not documented it, perhaps you should wonder what's going to happen when he walks in front of a truck, or is shot by an irate husband! In this case, none of the sysadmins (who are too busy) had any idea what was going on, and they didn't know enough about NTP to realize what was going on. The sysadmin had gotten the can't-sync error message many times, but didn't quite understand what it was saying. So even if he is hit by an irate truck, his replacement won't necessarily be better or worse. The problem I'm trying to solve is different. We put NTP on lots of different kinds of computer, mostly Unix, but some Windows, and I'm looking for diagnosis tools that will tell me what's really going on, precisely so I can debug unfamiliar setups no matter how screwed up. Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
[ntp:questions] Finding out where ntpd gets its ntp.conf file
We had been struggling with NTP running on Red Hat Enterprise Linux (RHEL) on IBM-built Intel boxes, specifically with getting NTP to generate loopstats and peerstats files. Basically, nothing worked, despite many attempts. Yesterday, I cracked it while working on the problem with one of the sysadmins. The ntp.config file was very complex, and I suspected it of being mostly flotsam and jetsam from prior uses, and probably in conflict with itself, so we cleaned it down to maybe three lines, and then stopped and started ntpd using the service utility. The daemon started, but complained that it was unable to synchronize. The sysadmin mentioned that it very often did this, for unknown reasons. Then I noticed that the timeserver IP address was not the same as specified in the simplified ntp.conf file, and sure enough the address that NTP was trying to use was not accessible to ping. Hmm. Huh? NTP cannot be using the ntp.conf file we thought it was. Tried starting NTP manually with -c option and providing the full path to our ntp.conf file. Success! Read the service shell script. It appears to get its file paths from environment variables named after the thing being started and stopped and accessible only in the root environment; this bit of RHEL-specific structure is being chased down. (Does anyone know where this is documented?) Which brings me to a question: How does one get NTP to tell you exactly where it is getting such things as the ntp.conf file from, all without being able to find or see the actual command line or lines that launched the daemon? I did not see a ntpq command that sounded plausible, although ntpq would be an obvious choice. This would be very useful for debugging, as each and every platform type seems to have a different approach to handling NTP. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] NTP Drifts +ve and -ve
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Arul Murugan) wrote: Hi, We are using NTP4, when CPU is very busy some of the UDP packets [are] dropped by the kernel, so the local clock drifts 60 milliseconds from the time server. Dropped packets are quite unlikely to be the problem, even if most packets never arrive. More likely is that the NTP daemon is being preempted between taking the send timestamp and the sent packet actually appearing on the wire, and between received packets actual arrival time and when the daemon is able to obtain the receipt timestamp. These preemptions appear to the daemon as very large and random asymmetrical transport delays. If sufficiently common, these bad observations will seep through the various filter steps in NTP, and corrupt the measurements of clock offset error used to update the servo. See http://www.eecis.udel.edu/~mills/stamp.html. What computer platform and operating system are you using? One classic solution is to give the NTP demon sufficient realtime priority to outrank whatever else the CPU is doing, thus sharply reducing fraction of NTP polls that suffer preemption. This raised priority will not cause those other activities to be any slower because the NTP daemon is an insignificant consumer of CPU resources. From that point, NTP keeps drifts +ve and -ve for 2 to 3 three days to become stable. The graph looks a like a sine wave oscillating and reaching zero after 3 days. My question are: 1. Why [is] NTP drifting +ve and -ve? Because the clock servo is being fed contaminated data, as explained above. 2. Why should NTP [be] taking 3 days for correcting 60 milliseconds? Because it takes NTP days versus a few hours to slog through all that bad data. 3. Is this a problem or it is expected? Both. It is a problem for sure, but is to be expected under these circumstances. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Dual Mixer Time Difference (DMTD) instruments sought
In article [EMAIL PROTECTED], jlevine [EMAIL PROTECTED] wrote: Hello, While it's unlikely that I will soon get to build such an instrument, I am quite interested in how they are built, if only to understand what can happen and why. Can you suggest some articles and/or books and/or patents delving into both the theory and the practicalities of building DMTD instruments? We (the time and frequency division of NBS/NIST) designed and built a dual-mixer systerm in 1980 (more or less). This same system is the one that still runs the atomic clock ensemble in Boulder. You can get the publications that describe this instrument from the publications database on our web site. Go to tf.nist.gov and click on the publications menu. When the menu appears, look for author Glaze. The stuff was published in about 1983 or so. There were several papers as I recall with various combinations of the folks who built the system and the software drivers for it. This is precisely the kind of pointer I was hoping for. Thanks. The system we built was totally analog, but a modern system would probably be fully digital. Our system had a resolution of about 0.2 ps and a stability of about 3-4 ps. A digital system could do better, mostly because the temperature sensitive stuff could be confined to the analog front end whereas we had to worry about temperature pretty much everywhere in the system. That isn't bad for 1980 analog electronics. I think that the 5120 is the digital realization, as discussed in other postings. That said, the 5120 is temperature sensitive, and one had to allow many hours for temperatures to stabilize, but then the resolution appeared to be about 0.01 pS. I assume that the improvement from 0.2 pS was due to the fancy matched-mixers trick, combined with use of a very low noise oscillator. However, the job is not trivial, since even tiny impedance mismatches can cause problems at this sub-picosecond resolution. You should watch especially for the connectors and the cables. We typically use SMA connectors and rigid coax. The inputs are buffered with distribution amplifiers with a reverse isolation that is as good as we can make it. About -165 db, I think, although I have not looked at that recently. (Note that the problems are not adequate digital computing power but plain old analog electronics.) As I said, I don't think I will be building such an instrument. But it's just this kind of nitty gritty detail I want to be aware of, for interest, and for self-protection in the lab. Even so, we have a detectable sensitivity to temperature at the level of ps. This noise level tends to be too small to affect the data from cesium standards, but it could be a problem if you were trying to calibrate the long-period performance of a device or a transmission system that had a small delay, since the residual diurnal temperature sensitivity could come to get you. What we were doing was to measure the temperature coefficient of electrical length of a temperature-stable 10 MHz distribution amplifier, the goal being a tempco not exceeding 1.0 pS per degree centigade. Some of the tested amps achieve ~0.5 pS/degree C, in a total delay of ~4.5 nanoseconds, or ~111 ppm per degree C, call it 100 ppm. The test consisted of measuring changes in total delay at three temperatures, 17, 24, and 31 degrees C. The problem is that it took at least an hour for the amplifier to stabilize at each temperature, so instrument drift is a significant source of error. The measured RC time constant of delay of the amplifier in chamber is 14 minutes. My solution was to compare the amplifier under test to a mechanical variable delay unit (Colby Instruments PDL-100A-625PS-5.0NS), using a fast sampling scope (200 femtosecond rms jitter(?), averaged down to ~50 fS) as the null detector. The specific circuit is a low-noise oscillator (Symmetricom 1050A) driving the first splitter, one output driving the scope sync input, the other driving the input of the second splitter. One output of the second splitter drives the reference path, which contains the variable delay unit. The other output drives the device path, which contains the amplifier under test. Both device and reference path cables pass through the environmental chamber, with the heated lengths held equal. The cables are low tempco as well (~1.5 ppm per degree C). Everything was 50-ohm, at least nominally, but no attempt at precision matching or isolation was made, and the connectors and adapters were a mix of whatever could be scrounged up in the lab. This setup yielded clean data, easily sufficient to the purpose. The main limits to accuracy appear to be hysteresis in the amplifiers under test, and the cyclic temperature variation of the environmental chamber itself. If you are in this business then you need professional help. Heh. I've been told this before, but the issue
Re: [ntp:questions] Dual Mixer Time Difference (DMTD) instruments sought
In article [EMAIL PROTECTED], Uwe Klein [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: OK. It sounds like what the 5120 does. I be that there are a lot of details to get *exactly* right, though. Right. But with having a ten year old Cray in every laptop ... Computational power must be harnessed to be useful. I'm talking about the considerable human effort required for the harnessing. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Dual Mixer Time Difference (DMTD) instruments sought
In article [EMAIL PROTECTED], jlevine [EMAIL PROTECTED] wrote: I may need a Dual Mixer Time Difference (DMTD) instrument, to measure picosecond changes in electrical length in a coax plus amplifier time reference signal distribution system with total delays in the hundreds of nanoseconds, currently operating at 10 MHz (sinewave), but with 100 MHz likely at some future date. What DMTD instruments are commercially available? A google search was not successful - all noise no detectable signal, probably because DMTD instruments are not that common, and many people build their own. We use dual-mixer systems in our primary time scale and also to calibrate and evaluate oscillators and timing hardware. So far as I know, the only units that are commercially available are made by Timing Solutions, which was recently acquired by Symmetricom. There are a number of different configurations, depending how how many devices you want to measure, whether they all run at the same frequency, etc. That's been what I'm finding, and now this is being confirmed. I don't know why Symmetricom keeps the 5120 under their hat. It's really a strange story - the only way to find out that the 5120 is a DMTD instrument (done up in all-digital DSP form) was by knowing that TSC used to make an analog DMTD instrument, and following TSC's (and specifically Dr Stein's) trail in the literature. It is possible to build these devices on your own, but it is not trivial to get pico-second resolution and stability. Almost everything is temperature sensitive at this level of resolution. I think such instruments are also sensitive to user mood. While it's unlikely that I will soon get to build such an instrument, I am quite interested in how they are built, if only to understand what can happen and why. Can you suggest some articles and/or books and/or patents delving into both the theory and the practicalities of building DMTD instruments? Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Dual Mixer Time Difference (DMTD) instruments sought
In article [EMAIL PROTECTED], David L. Mills [EMAIL PROTECTED] wrote: Joseph, I took a look at the instrument instruction manual to see what is going on. In typical todayspeak, Symmetricom doesn't say how the gadget works. I make it what used to be called a Costas direct-conversion receiver. The test signal is connected to two mixers; the reference oscillator is connected to the other mixer inputs in quadrature. The mixer outputs are digitized and filtered, the Q signal is shifted 90 degrees from the I signal and combinted. The result is a baseband SSB dignal which is then Fourier transformed for display. Is this what you have in mind? Yes, but not quite the whole story. Although impossible to discern from Symmetricom's 5120 datasheet and users guide, there is more to it than that. I found this instrument by accident while researching the literature for DMTD information. This search led me to Timing Solutions Corp (which was bought by Symmetricom in 2006) and Direct-Digital Phase-Noise Measurement, J. Grove, J. Hein, J. Retta, P. Schweiger, W. Solbrig, and S.R. Stein, 2004 IEEE International Ultrasonics, Ferroelectrics, and Frequency Control Joint 50th Anniversary Conference, pages 287-291. But if this is an advance in the technology, there could be a patent, and there was: Two-Channel Digital Phase Detector, US Patent 7,227,346 to Wayne E. Solbrig. I then approached Symmetricom, which led me to the 5120 (1 MHz to 30 MHz) and the 5125 (future, 1 MHz to 400 MHz). A section of the above article appears in the 5120 users guide. I have no idea why Symmetricom doesn't really mention that the 5120 can do these things, but I assume that the market for phase noise test sets vastly exceeds all other markets for a 5120-like instrument. I borrowed an early demo 5120 instrument, and in my somewhat slapdash lab setup, it was easily able to resolve 0.01 picosecond (eyeball rms width of the traces) changes in delay at 10 MHz while using a very quiet oscillator (a Symmetricom 1050A), after warming up overnight. Joe Gwinn Dave Joseph Gwinn wrote: In article [EMAIL PROTECTED], Joseph Gwinn [EMAIL PROTECTED] wrote: I may need a Dual Mixer Time Difference (DMTD) instrument, to measure picosecond changes in electrical length in a coax plus amplifier time reference signal distribution system with total delays in the hundreds of nanoseconds, currently operating at 10 MHz (sinewave), but with 100 MHz likely at some future date. What DMTD instruments are commercially available? A google search was not successful - all noise no detectable signal, probably because DMTD instruments are not that common, and many people build their own. The silence, the silence. I have not found too many commercial DMTF units, but I have found one, although the maker does not market it a such: The Symmetricom 5120 http://www.symmttm.com/products_pn_adev_test_sets_5120A.asp is at heart a digital DMTD instrument, and will make all the usual DMTD measurements, although it is marketed primarily as a phase noise test set. What else is available? Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Dual Mixer Time Difference (DMTD) instruments sought
In article [EMAIL PROTECTED], Uwe Klein [EMAIL PROTECTED] wrote: Joseph Gwinn wrote: I may need a Dual Mixer Time Difference (DMTD) instrument, to measure picosecond changes in electrical length in a coax plus amplifier time reference signal distribution system with total delays in the hundreds of nanoseconds, currently operating at 10 MHz (sinewave), but with 100 MHz likely at some future date. What DMTD instruments are commercially available? A google search was not successful - all noise no detectable signal, probably because DMTD instruments are not that common, and many people build their own. Thanks, Joe Gwinn Take one of the better GS DSO's that have high storage depth. Read the shots from the DSO and do all further processing in software? I don't understand how this would work. Could you expand the description? And what is GS? Thanks, Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Dual Mixer Time Difference (DMTD) instruments sought
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (John Ackermann N8UR) wrote: Joseph Gwinn said the following on 05/12/2008 10:38 PM: What DMTD instruments are commercially available? A google search was not successful - all noise no detectable signal, probably because DMTD instruments are not that common, and many people build their own. The silence, the silence. I have not found too many commercial DMTF units, but I have found one, although the maker does not market it a such: The Symmetricom 5120 http://www.symmttm.com/products_pn_adev_test_sets_5120A.asp is at heart a digital DMTD instrument, and will make all the usual DMTD measurements, although it is marketed primarily as a phase noise test set. What else is available? The 5120A is truly a wonderful box, but it's also not cheap (about $30K). It's fully DSP based so all the interesting stuff is done in software. One huge advantage is that the reference and device-under-test do not have to be at the same frequency. There's an older version, the 5110A, that has been discontinued but should sell used for less than $10K if you can find one. It's more of a pure DMTD box and doesn't do phase noise in a useful way. The 5110A is analog, I think, although I never did get a users guide. I don't know of other commercially marketed products that provide a DMTD function. However, there's been quite a bit of discussion about this over on the time-nuts list, and that's probably a better place for your question (https://www.febo.com/mailman/listinfo/time-nuts). I joined, but will lurk for now. The single most critical piece of a DMTD system is the zero crossing detector. Unless you have a way to increase the slew rate of the low frequency beat note by a million or so, trigger jitter in the counter will eat up almost all the advantages of the down-mix. Again, there's been some discussion about this on time-nuts, and there are some folks there working on designing and building bits of the hardware (at least, a couple of months ago there was a fair bit of discussion on the point). Yes. And don't forget ground loops. Noise at 1 Hz is very difficult to shield. I bet one big advantage of the DSP approach is that math is cleaner than practical analog hardware. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] no ntp synchronisation: 2s to 6s time shift !
In article [EMAIL PROTECTED], Richard B. Gilbert [EMAIL PROTECTED] wrote: Hans Jørgen Jakobsen wrote: On Sat, 23 Feb 2008 23:38:09 GMT, Danny Mayer wrote: No need to look. I haven't had the bandwidth to get much done in the last few months. Needless to say it will be much easier to implement on Windows than on Unix machines as Windows uses threads so there's no issue with the transfer of information between processes. Why are threads not used on UNIX? /hjj I've never used them but I believe that Solaris supports threads. True. IRIX (SGI) and AIX (IBM) also support threaded applications. The Solaris kernel is also itself threaded. I think this is true of IRIX and AIX. The HP-UX kernel is not threaded, although threaded applications are supported. These are the OSs I know something of. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Lep seconds
Dave, In article [EMAIL PROTECTED], David L. Mills [EMAIL PROTECTED] wrote: Joseph, Conversely, if a client syncrhonizes to a server strictly running TAI and never signals leaps, NTP will deliver TAI. NIST, USNO and I have discussed this serveral times and concluded the lessor of two evils is to continue with NTP on UTC. Yep. True enough. But GPS emits TAI (plus an offset), so one can claim that configuring the NTP timeserver to emit GPS System Time (not UTC) is to generate what is essentially TAI. This is widely done in the big-radar world. Joe Dave Joseph Gwinn wrote: In article [EMAIL PROTECTED], [EMAIL PROTECTED] (David Woolley) wrote: In article [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: compliant. Is there a similar mod for NTP. I am hoping that there is a mod that will cause NTP to supply theoretical UTC (even if it is not ascci). Both POSIX and NTP use UTC. Your problem is that you are not using using UTC, but, rather, using TAI. Actually, POSIX does *not* use UTC in the normal sense of the word, as no leap seconds are applied. The fundamental POSIX timescale counts what amount to SI seconds from the POSIX Epoch, 0h 0m 0s UTC 1 January 1970. Every day contains exactly 86,400 seconds. That said, if one drives a POSIX box via NTP from a GPS timeserver set to emit UTC (versus GPS System Time), time on the POSIX box will be pretty close to UTC. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] SNMP support
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Svein Skogen) wrote: (I took the liberty of removing the lower half of this mail. See previous mails in this thread for complete history) David L. Mills wrote: Heiko, A couple of comments about this mission. First, last I looked SNMP had a really hard time with floating point and the scaling issues are dangerous. Second, as mentioned several times on the NTP hackers wire, we would very much like to shoot ntpdc and its fascist (mode 7) protocol. As of now, many configuration issues can be performed using the mode-6 (ntpq) protocol. While many ntpdc related issues can be easily moved to the mode-6 protocol, which is based on UDP, the monlist function of ntpdc really needs TCP, as experience with monlist and UDP demonstrates. This paritcular combination of UDP and TCP would not be friendly to SNMP. I continue to speculate that an SNMP agent in an expert system would be an ideal shotgun marriage between mode-6 and SNMP. Dave Disclaimer: I know parts of this has already been answered in the thread, and I know that a lot of the basis for my comments are made solely based on my memory of things, and memory (when you start to get my age) isn't a perfect match of things that were, but rather some guidlines to how things might have been. Thus I may be totally wrong, or answering the wrong question. (Now, I can get to the point. :) ) One of the tricks I used in the old days for handling decimal numbers (which is why we need the floating point, isn't it?) was to use two variables, or to use a different (moving the decimal point) internal value, and dividing by 10^x for display. I'm guessing that what we need the floating point for, is the precision on our peers, and the precision of our drift. These values are (iirc) today a float number of milliseconds. And for all simplicity, they should remain that way for human presentation, to avoid unnecessary confusion. An alternative that comes to mind is to use the scaled binary representation of the base two logarithm of the millisecond value in question. Zero would need to be handled as a special case. If the value is signed, then a sign field will also be needed. For example, one could express 274591 milliseconds as 1000*Log2(274591)= 18066.9248, or rounded to the integer 18067. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Any samples for NTP/SNTP client code?
In article [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: Does anybody know of any *practical* samples on how to implement NTP/SNTP client?. The goal is to provide accurate time for a program/client running on Windows Vista. Specifically, what values to include in the the request message, how to process the reply message, etc. I am NOT asking how to send/receive UDP datagrams, or where to find comprehensive descriptions like RFC documents, or how to build or design user interfaces. Only a narrow description focused on NTP/SNTP request/reply datagrams for a simple PC client, preferably in C/C++ source code. I've done this in an embedded realtime system. (No, the source code is not available.) In Appendix A of RFC-1305 you will find the format of the NTPv3 request/response packet. Send this packet to port 123 of the NTP server, and read the reply packet. It's pretty easy. The NTPv3 packet format will work with all timeservers of NTPv3 and above. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] NTP architecture recommendation
In article [EMAIL PROTECTED], Richard B. Gilbert [EMAIL PROTECTED] wrote: David J Taylor wrote: Richard B. Gilbert wrote: [] I don't think that a 14 channel receiver would be useful! There simply are not that many satellites! The last I knew, there were 27 NavStar (GPS) satellites in orbit. Of these, about seven are usually above the horizon at any one time. Richard, There can be more than that - here's an example with 13 visible. http://www.david-taylor.myby.co.uk/software/wxtrack-extras.htm According to this source, there are 31 active satellites: http://celestrak.com/NORAD/elements/gps-ops.txt More in orbit, I expect, now dead. What do they do with decommissioned GPS satellites? Cheers, David I suspect there's not much they can do with decommissioned GPS satellites! You could probably launch 27 GPS satellites for the cost of one manned mission to retrieve/repair one. If you wait long enough, they will come down by themselves! It may take a few hundred years. . . . If I recall, they have a deorbiting system, which is a rocket that fires against their orbital motion, causing them to fall out of orbit rather more quickly than that. Of course, if the satellite has completely failed, the deorbit system won't be listening for commands, but usually things won't get that bad suddenly. Joe Gwinn ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions