| Re: AW: AW: [chrony-users] chronyd: Can' Synchronize WHY ? |
[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-users Archives
]
On Fri, 9 Sep 2011, thomas.schmid@xxxxxxxxx wrote:
Hi Bill,
Overlap
A says that the time offset is 3ms plus or minus 1 ms. Ie, it says the time is
between 2 and 4 ms out. B saus that the offset is -6ms plus or minus 2 ms, ie
from -8 to -4 ms out. Which of the two is the system to use?
Is the clock between -8 and -4ms out or is it between 2 and 4 ms out?
The two bands do not overlap. They are incompatible and chrony cannot decide.
To the point, thanks. I also read the article in about the "selection" mechanism in
"http://www.eecis.udel.edu/~mills/ntp/html/select.html"; quoted by Miroslav, which
(for NTP) was surprisingly easy to comprehend. I tend to treat NTP and its mechanisms
as a black box: "Believe it, it works, smart people have devised it". But of course
if the operating conditions change, this might no longer be true..
I am still trying to find out on why I have 3 (normally reliable) time sources which
deviate from each other in such a way, that no solution to whom might be right can
be drawn by chronyd.
Yes, and especially that the variance of the slope is SOOOOO huge.
Do you have measurements.log logged? Perhaps you could plot the meausured time
offset of the various sources vs time to see what in the world is happening.
I suspect that the fact that you are bouncing between sources so badly and
those sources disagree with each other so badly is what is causing that.
(Note sure I can convince myself that that could be the problem though)
On Fri, 9 Sep 2011, thomas.schmid@xxxxxxxxx wrote:
Hi Miroslav,
-----Ursprüngliche Nachricht-----
Von: Listengine [mailto:listengine@xxxxxxxxxxxxxxxxx] Im Auftrag von
Miroslav Lichvar
Gesendet: Donnerstag, 8. September 2011 10:58
An: chrony-users@xxxxxxxxxxxxxxxxxxxx
Betreff: Re: [chrony-users] chronyd: Can' Synchronize WHY ?
On Thu, Sep 08, 2011 at 06:35:34AM +0000, thomas.schmid@xxxxxxxxx
wrote:
^x <WAN-IP1> 1 6 49 +13us[ -
877us]
+/- 1148us
^x <WAN-IP2> 2 6 46 +192ms[
+192ms]
+/- 56ms
Yee gads, that's pretty horrible, except for wanIP1
IP2 seems really really far away. Why is it taking so long?
<WAN-IP1> is connected by a 7 Mbps WAN connection, 2 LAN and 1 WAN-
hop away.
<WAN-IP2>: The intermediate NTP server (WinXP+Meinberg ntpd) sits in
the same
LAN; however the WinXP is a virtual machine running on VMware Server
1.0.10.
Note: I know it's a big NO NO to run NTP servers out of (VMware)
virtual machines,
The intervals for the two sources don't overlap, and there is no third
source which would overlap one of the two, so chrony gives up on them.
Would you be so kind as to give me a "Dummie's Guid to "Overlap" ? I do
not understand much of the values presented (dropped out of too math class
too early :-)
If you want to keep the WAN-IP2 source as a backup, you can try to
increase its stratum by the minstratum option.
^x SU-Server1.<domain> 0 7 10y +0ns[ +0ns] +/-
0ns
^x LU-Server.<domain> 0 7 10y +0ns[ +0ns] +/-
0ns
Well something is certainly screwed up there. This says that the
last
measurement was 10 years ago. Also while this says it is 0 ns
jitter,
below it
says it is 4000ms jitter.
Something is seriously messed up here
These are just the default values when no samples were collected yet.
You only have two servers, the WAN servers, and they disagree on
what
the
right time is. Chrony has no idea which one to believe, since there
are
only
two of them (whichis why you should have an odd number.)
Hm, even if I have 3 valid NTP servers (setup with a valid <WAN-IP1>,
<WAN-IP2> and "SU-Server3") I see the "can't synchronize" message a
lot.
Is stratum for the third server 3 or higher? The sources output might
help.
I tested the 3 reference setup on another system: There are some "can't
synchronize" messages, less than the original 2 reference server setup,
but still a few:
Setup (chrony_server.conf, included into chrony.conf):
server <WAN-IP1> iburst prefer presend 5
server su-server3 maxpoll 7
server <WAN-IP2> presend 5
server su-server1 maxpoll 7
server lu-server maxpoll 7
initstepslew 30 <WAN-IP1> <WAN-IP2>
Again, su-server1 and lu-serevr are not reachable
syslog:
...
Sep 8 23:07:56 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:08:12 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:11:26 Com-PC chronyd[3237]: Can't synchronise: no majority
Sep 8 23:12:08 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:13:40 Com-PC chronyd[3237]: Selected source <WAN-IP2>
Sep 8 23:14:26 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:18:15 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:18:59 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:19:00 Com-PC chronyd[3237]: Can't synchronise: no majority
Sep 8 23:19:24 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:20:09 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:22:27 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:23:59 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:24:45 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:30:53 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:32:45 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:34:16 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:35:12 Com-PC chronyd[3237]: Can't synchronise: no majority
Sep 8 23:35:24 Com-PC chronyd[3237]: Selected source <WAN-IP2>
Sep 8 23:36:11 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:37:42 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:39:36 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:41:08 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:41:53 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:43:02 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:44:11 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:45:57 Com-PC chronyd[3237]: Can't synchronise: no majority
Sep 8 23:46:27 Com-PC chronyd[3237]: Selected source <WAN-IP1>
Sep 8 23:49:11 Com-PC chronyd[3237]: Selected source SU-Server3
Sep 8 23:49:54 Com-PC chronyd[3237]: Selected source <WAN-IP1>
...
chronyc:
chronyc> sources
210 Number of sources = 5
MS Name/IP address Stratum Poll LastRx Last sample
============================================================================
^* <WAN-IP1> 1 6 67 -36us[ -463ms] +/- 85ms
^+ SU-Server3 2 6 42 +394ms[ +394ms] +/- 133ms
^+ <WAN-IP2> 2 6 35 +83ms[ +83ms] +/- 47ms
^? SU-Server1 0 7 10y +0ns[ +0ns] +/- 0ns
^? LU-Server 0 7 10y +0ns[ +0ns] +/- 0ns
chronyc> sourcestats
210 Number of sources = 5
Name/IP Address NP NR Span Frequency Freq Skew Offset Std Dev
==============================================================================
<WAN-IP1> 7 5 412 -1555.346 3022.979 -9053us 2933us
SU-Server3 4 4 193 -2718.528 18920.604 +323ms 873us
<WAN-IP2> 5 4 342 -2080.739 9921.691 +56ms 5675us
SU-Server1 0 0 0 0.000 2000.000 +0ns 4000ms
LU-Server 0 0 0 0.000 2000.000 +0ns 4000ms
I have attached the (I think) relevant log files as well as the
sources/sourcestats/tracking output of chronyc.
Regs,
Thomas Schmid
--
William G. Unruh | Canadian Institute for| Tel: +1(604)822-3273
Physics&Astronomy | Advanced Research | Fax: +1(604)822-5324
UBC, Vancouver,BC | Program in Cosmology | unruh@xxxxxxxxxxxxxx
Canada V6T 1Z1 | and Gravity | www.theory.physics.ubc.ca/
---
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.
--
William G. Unruh | Canadian Institute for| Tel: +1(604)822-3273
Physics&Astronomy | Advanced Research | Fax: +1(604)822-5324
UBC, Vancouver,BC | Program in Cosmology | unruh@xxxxxxxxxxxxxx
Canada V6T 1Z1 | and Gravity | www.theory.physics.ubc.ca/