[ntp:questions] NTP not working after server crash
Hi, I have a server running VMWare Server, which contains a number of virtual machines. The host VM and guest VM's are all running CentOS 5 64bit. I installed NTP about a year ago and got it working on all of the virtual machines on this server, and it was working fine. However, I've just noticed that the time on one of my virtual servers is out, but it is still working fine on all of the other virtual machines. I checked the NTP log file on the virtual machine that is having problems and it just stopped doing anything in April: 16 Apr 12:34:49 ntpd[680]: no servers reachable 16 Apr 12:43:28 ntpd[680]: synchronized to 192.168.15.1, stratum 3 16 Apr 12:49:45 ntpd[680]: time reset +119.777502 s 16 Apr 12:55:03 ntpd[680]: synchronized to 192.168.15.1, stratum 3 16 Apr 13:30:30 ntpd[680]: time reset -0.350146 s 16 Apr 13:34:54 ntpd[680]: synchronized to 192.168.15.1, stratum 3 16 Apr 14:39:21 ntpd[680]: no servers reachable 16 Apr 14:50:03 ntpd[680]: synchronized to 192.168.15.1, stratum 3 12 Aug 16:26:08 ntpd[2187]: ntpd exiting on signal 15 12 Aug 17:52:39 ntpd[2440]: ntpd exiting on signal 15 12 Aug 20:35:29 ntpd[3009]: ntpd exiting on signal 15 I have verified that NTP is still running, and as you can see by the last three lines it is still logging to the log file when I restart it. My memory is a little sketchy, but I believe that on April 16th (the date NTP stopped working on this virtual machine), VMWare crashed and I had to do a forced reboot of the server (along with all the VM's). I'm not sure how or why this would cause the problem, but I assume that it is related. Here is my /etc/ntp.conf file: # ntpd.conf tinker panic 0 restrict 127.0.0.1 #restrict default notrust nomodify nopeer server 192.168.15.1 driftfile /var/lib/ntp/drift logfile /var/log/ntp.log This config file is completely identical to all of the other virtual machines that do work. I have verified that the file permissions on /var/lib/ntp/drift are correct and I have confirmed that there is no firewall causing problems. Running ntpq -p on the broken VM returns: remote refid st t when poll reach delay offset jitter == 192.168.15.1.INIT. 16 u- 6400.000 0.000 0.000 And on one of the other VM's that is working: remote refid st t when poll reach delay offset jitter == *192.168.15.1173.201.38.853 u 886 1024 3771.263 28.732 57.120 I have confirmed that I can ping 192.168.15.1 (which is the internal IP address for the host machine) from the VM that isn't working. Does anyone know what could be wrong? Cheers, Andrew ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
[ntp:questions] running NTP as server only
Hi, Is it possible to run the NTP daemon only as a server and not as a local-clock maintainer? Reason: I have a virtual machine which gets its time via the vmware tooling from the hardware server it is running on. Now this virtual machine needs to distribute the time to clients. Folkert van Heusden -- Multitail es una herramienta flexible que permite visualizar los log file y seguir la ejecución de comandos. Permite filtrar, añadir colores, combinar archivos, la visualización de diferencias (diff- view), etc. http://www.vanheusden.com/multitail/ -- Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] use ntpd, how to know unsync status on linuxsystem
I`m using ntpd(ntp4.2.6p1) as client in linux(2.6.26). I have some questions about my user app how to know the ntp client sync status. 1. I use ntp_gettime, is there other way? That depends on what you want to know about the state of ntpd. You probably should want to know the error bound, in which case it is an appropriate method. If you really want to know whether it is synched, you need to use the NTP protocol to either check the stratum (note that this may take a long time to go to 16, when the last server stops responding, or more detailed state variables. Thank you, David. I just forcus on the return value of ntp_getime, but don`t know the stratum. I think the stratum means the time quality, is right? 2. If there is no server alive, when I start ntp client, why ntptime show :ntp_gettime() returns code 0, but not code 5? I believe there was a policy decision some time ago never to declare a failure to this interface. I trace the function loop_config in the file ntpd/ntp_loopfilter.c : I think it should be : ntv.maxerror = (u_int32)(MAXDISPERSE * 1e6); ntv.esterror = (u_int32)(MAXDISPERSE * 1e6); I tend to agree. or not? 3. If ntp server no answer for 1h or shortter time, ntpd can mark unsync stat? how could I config ntpd? Why is this important to you? The time should still be very accurate at that point. My project is about receive sample value for the merge unit, by IEC61850-9-2. The merge unit (MU) only show the sample count in one second. eg: 0 -- 3999. So, my PC must get the SOC and associated with sample value. When the ntpd, which run on my PC as client, cann`t receive the reply from ntp server, the SOC got from my PC clock will offset about 10s in 24h. So, I`m apt to mark the unsync stat for the sample value so qickly. this is app map: ntpd --- linux system my app --- MU So, I realy want to know the unsync, as fast as I can, by some API. By your answer, I`ll try to check the stratum, and use local server 127...stratum 16. When remote server no reply received, can ntpd use the local server`s stratum? Maybe it help to change the stratum so qickly as I wash? When I waited the help, I try to write a simple sntp client for my project. When I use ntp_adjtime to adjust the offset, I see ntptime show: ntp_adjtime() returns code 0 (OK) modes 0x0 (), offset -535.000 us, frequency -500.000 ppm, interval 1 s, maximum error 42516 us, estimated error 16 us, status 0x1 (PLL), time constant 2, precision 1.000 us, tolerance 500 ppm, next to change : ntp_adjtime() returns code 0 (OK) modes 0x0 (), offset 0.000 us, frequency -500.000 ppm, interval 1 s, maximum error 454516 us, estimated error 16 us, status 0x1 (PLL), time constant 2, precision 1.000 us, tolerance 500 ppm, my function is : static inline __attribute__((always_inline)) int sntp_adjtime(double offset) { struct timex ntpx; ntpx.modes = 0; ntp_adjtime(ntpx); if ((ntpx.statusSTA_UNSYNC) || (fabs(offset) = 0.128)){ ntpx.modes = MOD_STATUS|MOD_MAXERROR|MOD_ESTERROR|ADJ_OFFSET; ntpx.status = STA_PLL; ntpx.maxerror = MAXDISPERSE; ntpx.esterror = MAXDISPERSE; ntpx.offset = (uint32_t)(offset*1e6); ntp_adjtime(ntpx); printf(sntp adjtime %f\n, offset); } return 0; } I don`t know the particular about how to use the ntp_adjtime. I want to know how the ntpd adjust the local clock, or where can I get the doc about ntpd work logic? ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions 本电子邮件及附件都属机密文件,并且仅供标明地址的个人或团体使用。 如果本电子邮件与您无关,请删除全部原始信息(包括任何原始信息的备份),并通知发件人。 This email and its attachments may be confidential and are intended solely for the use of the individual to whom it is addressed. If you are not the intended recipient of this email and its attachments, you must take no action based upon them, nor must you copy or show them to anyone. Please contact the sender if you believe you have received this email in error. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
[ntp:questions] phantom server with MacOSX
I've stumbled on a weird problem with NTP 4.2.6 on MacOSX. I compiled and installed the code myself. The box is not running whatever NTP code is shipped with the OS. All is well (sort of). The server is chiming away quite happily. However it's peering with an Apple NTP server that I didn't configure. Can anyone suggest what's caused that or how to fix it? This phantom NTP server persists across restarts of ntpd and reboots, even though it's not listed in /etc/ntp.conf. I have no idea where it's picking up the name of this server from = how to get rid of it. Although this is no big deal, I am puzzled and a bit annoyed that Apple appear to be stamping over my NTP setup and not explaining how or why they do that. It's the thin edge of the wedge. If His Steveness doesn't like my NTP configuration, what else might he take exception to? My guess is there's something buried deep in Apple's DirectoryServer. But even if there is some secret NTP stuff there, I don't understand how it's getting picked up by my own instance of ntpd. Unless perhaps it's using one of Apple's shared libraries which pulls in the DirectoryServer stuff by default. Any suggestions? Here's what's in /etc/ntp.conf # # Clocks for {gromit,wallace}.rfc1035.com # server shaun.rfc1035.com # 93.186.33.42 server hutch.rfc1035.com # 195.54.233.70 server uchiawase.rfc1035.com # 90.155.4.149 And here's what ntpq is reporting for my ntp server: gromit% ntpq ntpq peers remote refid st t when poll reach delay offset jitter == -time6.apple.com 17.254.0.49 2 u 91 128 377 231.236 -15.714 1.668 +shaun.rfc1035.c 192.36.133.252 u 103 128 377 52.237 -7.862 1.401 +hutch.rfc1035.c 195.66.241.102 u 12 128 3770.104 -7.524 1.367 *uchiawase.6uk.o 157.44.176.4 2 u7 128 377 53.420 -3.990 2.236 LOCAL(0).LOCL. 6 l 1733 6400.0000.000 0.000 ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] running NTP as server only
folkert folk...@vanheusden.com wrote: Hi, Is it possible to run the NTP daemon only as a server and not as a local-clock maintainer? Reason: I have a virtual machine which gets its time via the vmware tooling from the hardware server it is running on. Now this virtual machine needs to distribute the time to clients. Did you study the vmware technical note about keeping time in virtual machines? It suggests that turning off this time synchronization in the vmware tools and running normal ntp in the virtual machine is better... ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] running NTP as server only
Hi-- On Aug 17, 2010, at 7:58 AM, folkert wrote: Is it possible to run the NTP daemon only as a server and not as a local-clock maintainer? Reason: I have a virtual machine which gets its time via the vmware tooling from the hardware server it is running on. Now this virtual machine needs to distribute the time to clients. Virtual machines make terrible timesources-- 10's to 100's of milliseconds of jitter are not unusual. If you need to run ntpd on that specific hardware, run it in the host ESX or Xen's Dom0 instead, and not in one of the hosted virtual machines. Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] phantom server with MacOSX
Hi, RFC-1035-- On Aug 17, 2010, at 4:32 AM, ntpquesti...@rfc1035.com wrote: I've stumbled on a weird problem with NTP 4.2.6 on MacOSX. I compiled and installed the code myself. The box is not running whatever NTP code is shipped with the OS. All is well (sort of). The server is chiming away quite happily. You didn't mention which version of MacOS X you were running (uname -a or sw_vers). The default config under 10.6 looks something like: % ps auxww | grep ntpd | grep -v grep root55 0.0 0.0 2435208 1144 ?? Ss5Aug10 0:52.39 /usr/sbin/ntpd -c /private/etc/ntp-restrict.conf -n -g -p /var/run/ntpd.pid -f /var/db/ntp.drift % cat /private/etc/ntp-restrict.conf # Access restrictions documented in ntp.conf(5) and # http://support.ntp.org/bin/view/Support/AccessRestrictions # Limit network machines to time queries only restrict default kod nomodify notrap nopeer noquery restrict -6 default kod nomodify notrap nopeer noquery # localhost is unrestricted restrict 127.0.0.1 restrict -6 ::1 includefile /private/etc/ntp.conf % cat /private/etc/ntp.conf server time.apple.com However it's peering with an Apple NTP server that I didn't configure. Can anyone suggest what's caused that or how to fix it? This phantom NTP server persists across restarts of ntpd and reboots, even though it's not listed in /etc/ntp.conf. I have no idea where it's picking up the name of this server from = how to get rid of it. Pay attention to the launch arguments of ntpd; make sure you are running the binary you expect and pointing to the config file you expect. If need be, run ntpd -d, make sure it returns your 4.2.6 version rather than something like ntpd 4.2@1.1520-o which would be the base OS binary; and pay attention to newpeer events. If that's still not enough, dtruss /usr/sbin/ntpd -n -d (or whatever the path to your binary is) should return output like: write_nocancel(0x1, addto_syslog: Listening on routing socket on fd #27 for interface updates\n\0, 0x4A) = 74 0 write_nocancel(0x1, local_clock: time 0 offset 0.00 freq 0.000 state 0\n\0, 0x37) = 55 0 open_nocancel(/etc/ntp.conf\0, 0x0, 0x1B6) = 3 0 fstat64(0x3, 0x7FFF5FBFED60, 0x7FFF5FBFEE2C) = 0 0 read_nocancel(0x3, server time.apple.com\n\0, 0x1000) = 22 0 ...and that will show you exactly which config file is being opened and what was read from it. Although this is no big deal, I am puzzled and a bit annoyed that Apple appear to be stamping over my NTP setup and not explaining how or why they do that. It's the thin edge of the wedge. If His Steveness doesn't like my NTP configuration, what else might he take exception to? I admire the humor behind the notion that Steve would care about your NTP configuration; if the issue comes down to a user configuration issue, you'll hopefully retain this sense of humor. :-) Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] running NTP as server only
folkert folk...@vanheusden.com wrote in message news:20100817145830.gh5...@belle.intranet.vanheusden.com... Is it possible to run the NTP daemon only as a server and not as a local-clock maintainer? Reason: I have a virtual machine which gets its time via the vmware tooling from the hardware server it is running on. Now this virtual machine needs to distribute the time to clients. server 127.127.1.0 comes to mind. Groetjes, Maarten Wiltink ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] use ntpd, how to know unsync status on linuxsystem
ÕÔÓ±¿Æ wrote: I just forcus on the return value of ntp_getime, but don`t know the stratum. I think the stratum means the time quality, is right? http://en.wikipedia.org/wiki/Network_Time_Protocol#Clock_strata http://www.eecis.udel.edu/~mills/ntp/html/index.html -- 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] use ntpd, how to know unsync status on linuxsystem
ÕÔÓ±¿Æ wrote: I want to know how the ntpd adjust the local clock, or where can I get the doc about ntpd work logic? http://www.ietf.org/rfc/rfc5905.txt http://www.eecis.udel.edu/~mills/ntp/html/index.html http://www.ntp.org/downloads.html http://support.ntp.org/bin/view/Main/SoftwareDownloads -- 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] running NTP as server only
Hi, Is it possible to run the NTP daemon only as a server and not as a local-clock maintainer? Reason: I have a virtual machine which gets its time via the vmware tooling from the hardware server it is running on. Now this virtual machine needs to distribute the time to clients. Virtual machines make terrible timesources-- 10's to 100's of milliseconds of jitter are not unusual. I don't think that is in all situations the case. Depends on the scheduling by the hypervisor. iirc ibm pseries lpars don't have this problem. That is why they have (x)ntp running in each of them normally. If you need to run ntpd on that specific hardware, run it in the host ESX or Xen's Dom0 instead, and not in one of the hosted virtual machines. Yes, that is what I propose: - let the hypervisor sync to a reliable accure timesource - sync the vms to the hypervisor with some mechanism. e.g. on vmware you have the vmware tooling which runs in the vm and syncs time to the hypervisor (and also things like memory ballooning etc) - let the vm then distribute the notion of time it got from the hypervisor to clients Somewhere this week I'll test how this works: I put together what I was asking, a program which picks the time from the local clock and then sends this via ntp. Then I'll have two systems (which run directly on hardware, not a vm) that will have a couple of low-stratum servers to sync against as well as my vm. If they then declare my solution as a falseticker and/or with a high jitter, I then know it won't works. Agree? Folkert van Heusden -- Ever wonder what is out there? Any alien races? Then please support the s...@home project: setiathome.ssl.berkeley.edu -- Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] running NTP as server only
Hi-- On Aug 18, 2010, at 2:01 PM, folkert wrote: Virtual machines make terrible timesources-- 10's to 100's of milliseconds of jitter are not unusual. I don't think that is in all situations the case. Depends on the scheduling by the hypervisor. It depends on the hypervisor, the hardware, the workload, etc. iirc ibm pseries lpars don't have this problem. That is why they have (x)ntp running in each of them normally. The type of latency you might see from a lightly loaded box might well be a lot better than what you see from a virtual webhosting company which jams as many VMs as it possibly can before the clients yell about slow web performance... If you need to run ntpd on that specific hardware, run it in the host ESX or Xen's Dom0 instead, and not in one of the hosted virtual machines. Yes, that is what I propose: - let the hypervisor sync to a reliable accure timesource - sync the vms to the hypervisor with some mechanism. e.g. on vmware you have the vmware tooling which runs in the vm and syncs time to the hypervisor (and also things like memory ballooning etc) - let the vm then distribute the notion of time it got from the hypervisor to clients I agree up to the last part. The quality of time you get from querying ntpd running in the hypervisor should be noticably / obviously better than what you would get from ntpd running in a VM. Set up both and compare the jitter your clients see for yourself. Somewhere this week I'll test how this works: I put together what I was asking, a program which picks the time from the local clock and then sends this via ntp. Then I'll have two systems (which run directly on hardware, not a vm) that will have a couple of low-stratum servers to sync against as well as my vm. If they then declare my solution as a falseticker and/or with a high jitter, I then know it won't works. Agree? Yes, testing is a fine idea. Regards, -- -Chuck ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] phantom server with MacOSX
In article 1657.1282044...@gromit.rfc1035.com, wrote: Any suggestions? Chuck gave you the A answer. The short-cut answer, at least in 10.6, is: System Preferences - Date Time UNcheck Set date and time automatically. Your Mac will now stop infecting, er, injecting time.apple.com into your config file. I'm actually impressed, being fairly new to the Mac, that they're using NTP, instead of some Apple created utility, to set the time. And remember: Steve loves you -- Seth Goodman ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions