I configured UDel ntpd (version 4.2.6.p5) as a client to three servers operated by my organization, and (I thought) no one else.
server xx.yy.zz.1 iburst
server xx.yy.zz.2 iburst
server xx.yy.zz.3 iburst
restrict default ignore
restrict xx.yy.zz.1 nomodify notrap nopeer noquery
restrict xx.yy.zz.2 nomodify notrap nopeer noquery
restrict xx.yy.zz.3 nomodify notrap nopeer noquery
After about five minutes of uptime, ntpq reported that the daemon had discovered and added two more time servers (also operated by my organization, but not on their Official List of Time Servers):
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
*xx.yy.zz.1 aa.bb.cc.dd 2 u 16 64 1 0.157 -2.230 0.032
xx.yy.zz.2 aa.bb.cc.dd 2 u 15 64 1 0.207 -2.191 0.032
xx.yy.zz.3 aa.bb.cc.dd 2 u 14 64 1 0.211 -2.171 0.014
xx.yy.zz.4 .INIT. 16 u - 64 0 0.000 0.000 0.000
xx.yy.zz.5 .INIT. 16 u - 64 0 0.000 0.000 0.000
Since these servers were not in the address whitelist (because I didn't know about them), the client's attempt to communicate with them obviously failed.
I don't know exactly how this happened, but I get the impression from the documentation that there's some way for Server A to tell the client "hey, you should be talking to Servers B, C, and D too." The documentation I've found has a lot of rambly verbiage about the various methods for this, but absolutely nothing about how to control it.
Therefore, the question: How do I completely turn this off, so that the client only attempts to talk to the servers listed explicitly in the configuration file, no matter what?
Per request:
# ntpq -pncrv | redact
remote refid st t when poll reach delay offset jitter
==============================================================================
+xx.yy.zz.1 aa.bb.cc.dd 2 u 56 1024 377 0.177 -0.010 0.070
*xx.yy.zz.2 aa.bb.cc.dd 2 u 690 1024 377 0.237 0.028 0.046
+xx.yy.zz.3 aa.bb.cc.dd 2 u 226 1024 377 0.229 0.013 0.052
xx.yy.zz.4 .INIT. 16 u - 1024 0 0.000 0.000 0.000
xx.yy.zz.5 .INIT. 16 u - 1024 0 0.000 0.000 0.000
associd=0 status=061b leap_none, sync_ntp, 1 event, leap_event,
version="ntpd [email protected] Fri Apr 10 19:04:04 UTC 2015 (1)",
processor="x86_64", system="Linux/3.16.0-4-amd64", leap=00, stratum=3,
precision=-22, rootdelay=0.405, rootdisp=38.394, refid=128.2.1.21,
reftime=d951542d.a6db3cdc Wed, Jul 15 2015 17:50:37.651,
clock=d95156df.0d2756da Wed, Jul 15 2015 18:02:07.051, peer=9102, tc=10,
mintc=3, offset=0.009, frequency=-5.266, sys_jitter=0.024,
clk_jitter=0.030, clk_wander=0.003
# redact < /etc/ntp.conf
driftfile /var/lib/ntp/ntp.drift
statsdir /var/log/ntpstats/
statistics loopstats peerstats clockstats
filegen loopstats file loopstats type day enable
filegen peerstats file peerstats type day enable
filegen clockstats file clockstats type day enable
server aa.bb.cc.1 iburst
server aa.bb.cc.2 iburst
server aa.bb.cc.3 iburst
restrict -4 default ignore
restrict -6 default ignore
restrict aa.bb.cc.1 nomodify notrap nopeer noquery
restrict aa.bb.cc.2 nomodify notrap nopeer noquery
restrict aa.bb.cc.3 nomodify notrap nopeer noquery
restrict 127.0.0.1
restrict ::1
I don't see anything in here about picking up additional servers from DHCP.
I took a guess what your organization is and that you are using dhcp.
Yes your organization prominently publishes a group of 3 servers That being said your organization also instructs dhcp clients to use six other servers.
I am not sure why it is a problem that your machine is using servers advertised in dhcp but you can turn this functionality off if you want. If you are using Debian you need to edit
/etc/dhcp/dhclient.conf
and remove the ntp-servers option from the request statement:If you are not using Debian you need to consult your distros documentation.