[KLUG Members] ntpd and samba

[Adam Williams]

>>So when I start ntpd on the client machine, I get....
>>ntpd: Synchronizing with time server:                      [FAILED]
>>The log on the server is untouched, the log on the client is...
>>ntpdate[4029]: no server suitable for synchronization found
>>ntpd:  failed

Right.  The server clock either differs more than 1000 seconds from the
client, in which case NTP aborts with the long technical version of
"Your screwed"; or the server has not yet decreased its stratum to a
legitimate value.  Servers seem to maintain a ludicrous stratum, making
clients allergic to them, until they have been going long enough to be
fairly certain about their time value.

>How did you test that your server's ntpd starts serving time right away?

[root@littleboy /root]# ntpq
ntpq> pe
     remote           refid      st t when poll reach   delay   offset 
jitter
==============================================================================
+BITSY.MIT.EDU   GPS-1.MIT.EDU    2 u  818 1024  377   56.620   -4.436  
0.320
-bonehed.lcs.mit .GPS.            1 u  722 1024  377   70.479    1.286  
1.309
+clock.isc.org   gps.laguna.vix.  2 u 1005 1024  377   82.099   -1.999  
0.042
*darkcity.cerias .GPS.            1 u  575 1024  177   44.565   -3.374  
0.041
-ntp-0.gw.uiuc.e truechimer.cso.  2 u  738 1024  377   36.562    3.125  
0.531
ntpq>

When pe looks something like the above you should be good to go.

>Ntpdate and ntpd are two very different programs: your ntpd start script
>seems to call ntpdate first before it runs ntpd, and doesn't call it if
>ntpdate fails. Try running 'ntpdate <server>' from the client command
>line after your server's ntpd has been running for an hour or two.

Do a "ntpdate -d {server}" and what do you see?