[ntp:hackers] Unexpected origin timestamp

Havard Eidnes he at uninett.no
Wed Apr 20 09:17:15 UTC 2016


> It used to be that this situation caused a counter increment and was
> otherwise silent.
>
> This is the first time I've seen a report of the situation you describe.
>
> Is there anything interesting in the syslog on the sending system?

Not from ntpd, no.  Since both are configured with "peer other-address",
I now get the same log messages on both:

Apr 20 09:06:13 oliven ntpd[17457]: receive: Unexpected origin timestamp 0xdabfe612.67e4fee0 from 128.39.2.22 xmt 0xdac1a8e5.910fb80f
Apr 20 09:23:40 oliven ntpd[17457]: receive: Unexpected origin timestamp 0xdabfe612.67e4fee0 from 128.39.2.22 xmt 0xdac1acfc.9164c5a1

vs.

Apr 20 09:00:49 hegre ntpd[28988]: receive: Unexpected origin timestamp 0xdabfe619.904ed55b from 158.38.0.168 xmt 0xdac1a7a1.691d2331
Apr 20 09:18:00 hegre ntpd[28988]: receive: Unexpected origin timestamp 0xdabfe619.904ed55b from 158.38.0.168 xmt 0xdac1aba8.69d77b8b

As long as one of them ran the earlier version, the association worked
fine.

> What does 'ntpq -p' show on both machines?

oliven: {3} ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
+time2.stupi.se  .PPS.            1 u  939 1024  377   15.069    0.246   0.134
-ntps1-0.eecsit. .PPS.            1 u  937 1024  377   41.416   -6.694   0.137
+ntp1.nordu.net  192.36.144.23    2 u 1030 1024  377   14.273    0.200   0.136
+ntp1.sth.netnod .PPS.            1 u  238 1024  377   14.455    0.059   0.110
+77.40.226.114   .PPS.            1 u  143 1024  377    9.191    0.226   0.182
 hegre.uninett.n 194.58.202.20    2 u  34h  256    0    7.979    0.753   0.000
*ravnkloa.urc.un .PPS.            1 u   40 1024  377    0.719    0.106   0.092
oliven: {4}

hegre% ntpq -c pe
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 oliven.uninett. 158.38.39.5      2 s  34h   64    0    7.762   -0.861   0.000
-alesund-mp.hial 81.175.5.182     2 u  226  256  377   11.054    0.120   0.029
+ntp.uit.no      .GPS.            1 u  215  256  377   22.172    0.045   0.078
-pil32-mp.hioa.n 139.112.153.38   3 u  212  256  377    0.336   -1.242   0.157
-n1.taur.dk      .PPS.            1 u   99  256  377   16.347    0.497   0.046
*sth1.ntp.se     .PPS.            1 u  235  256  377    6.784    0.095   0.038
+ravnkloa.urc.un .PPS.            1 u  208  256  377    8.403    0.088   0.058
hegre% 

i.e. they have both stopped listening to each othes updates, but other
associations configured with "server xx.xx.xx.xx" work fine.

> There are some patches around peer mode in 4.2.8p7, scheduled for public
> release on Tuesday the 26th of April.  I have no idea if they will fix
> the situation you are seeing or not.  The peer mode tests have been
> working fine for me and several others who have been testing it.

Hmm, ok.

Regards,

- Havard


More information about the hackers mailing list