[ntp:questions] Re: Cannot synchronize to server with local clock

Tim tim at mail.localhost.invalid
Thu Jul 21 02:26:02 UTC 2005


Tim wrote:
 
>> My server was off-line for so long each day that it's stratum would drop
>> to 16 for most of the time, making it useless as a reference for my LAN
>> (they'll ignore it).
 

Serge Bets wrote

> Server's stratum is not supposed to raise to 16, because LOCAL() driver
> artificially extends sync forever. Monthes, years, eons. Show us
> experimented ntp.conf, please, there was surely a problem.

I'm not sure if I've fiddled with the configuration file since I ran the
experiment, and I know the machine's reconfigured the clients, itself,
several times.  I'll include the files, anyway.  Basically, the changes I
made were to what I thought I had to do to allow local machines to see the
server, and for the local machines to use the server.
 
>> the computers keep time well enough, just free wheeling along
>> individually, that they could carry on as they were until the next time
>> the LAN was connected to the internet.

> So you removed LOCAL()?

No.  What happened was they were ignoring my server, and doing whatever
they normally did when they didn't have access to a time server.

What I did to set up my experiment was configure *them* only with my local
time server address.  However, they'd always reconfigure themselves back
to having some other NTP servers, too.

This is playing with Fedora Core 3 Linux on those machines, with Red Hat
9.0 Linux on the server, by the way.

> My advice for you would be to set a 127.127.1.0 stratum 8 prefer on your
> server, and another one stratum 14 (but not prefer) on a given client.
> This way you have eternal sync, only one sync source at a time, and you
> can shutdown server *or* given client for maintenance.

I'll have another bash at things.  This is just for the sake of learning,
by the way.  Since I've found the machines to look after themselves rather
well, without my interference.  ;-)

> The danger here is sync source hopping and peeks in frequency it can give.
> Ntpd is not perfect for your dialup case. People say Chrony is better
> suited?

Not one that I've heard of.  Reading <http://chrony.sunsite.dk/> about it,
now...

---------- server configuration file (minus default comments) ------------

restrict default ignore
restrict 209.132.176.4 mask 255.255.255.255 nomodify notrap noquery
broadcast 192.168.1.0
# the following line added by me:
restrict 192.168.1.0 mask 255.255.255.0 notrust nomodify notrap
server 209.132.176.4
fudge 127.127.1.0 stratum 10
driftfile /etc/ntp/drift
broadcastdelay  0.008
authenticate yes
keys /etc/ntp/keys

---------- client configuration file (minus default comments) ------------

restrict default nomodify notrap noquery
restrict 127.0.0.1
fudge 127.127.1.0 stratum 10
driftfile /var/lib/ntp/drift
broadcastdelay  0.008
keys /etc/ntp/keys
# the following three lines get added by themselves, or ones like them
# even if I delete them.  It always went out and found more servers.
restrict 80.85.129.2 mask 255.255.255.255 nomodify notrap noquery
restrict 80.74.64.1 mask 255.255.255.255 nomodify notrap noquery
restrict 200.218.160.160 mask 255.255.255.255 nomodify notrap noquery
restrict 192.168.1.2 mask 255.255.255.255 nomodify notrap noquery
server 192.168.1.2
broadcastclient

-- 
If you insist on e-mailing me, use the reply-to address (it's real but
temporary).  But please reply to the group, like you're supposed to.

This message was sent without a virus, please delete some files yourself.




More information about the questions mailing list