On 2007-02-09, sandy <[EMAIL PROTECTED]> wrote: > I have problem when using SUSE Linux as time server. When I try to use > linux as time server, Linux replies with LI leap indicator flags set > to "alarm condition (clock not synchronized)". As per RFC 2030, as > unicast mode is used, as LI is 3, the reply is getting rejected. My > Linux time server is standalone system. It is not getting time synched > with external source. Could this the be the reason ?
ntpd must be synchronized to _something_, even if it's just the Undisciplined Local Clock, or LocalCLK, (i.e. 127.127.1.x), before it can serve time to others. The best source of time is a ref-clock or another time server which is (ultimately) synced to a ref-clock. You may use remote time servers from the NTP Pool (http://ntp.isc.org/pool or http://www.pool.ntp.org/) or select servers from the Stratum-2 list (http://ntp.isc.org/s2 or http://www.ntp.org/s2). You should use 4, or more, remote time servers. You can use only 1 or 3, although that does not provide enough redundancy). Do not, under any circumstances, use 2. If your ntpd is not able to poll a real time source (either a local ref-clock or a remote time server) then you will have to configure the LocalCLK as follows: server 127.127.1.0 fudge 127.127.1.0 stratum 10 Then restart ntpd and wait ~5 to 8 minutes. Then your SNTP clients should accept the time from your ntpd. You should see the following 'ntpq -p' billboard when ntpd is synced to only the LocalCLK: remote refid st t when poll reach delay offset jitter ================================================================= *LOCAL(0) .LOCL. 11 l NN NN 377 0.000 0.000 0.004 If are unable to synchronize to your ntpd, please post the following: * Your 'ntpq -p' billboard * Your /etc/ntp.conf file (without the comment lines) -- Steve Kostecke <[EMAIL PROTECTED]> NTP Public Services Project - http://ntp.isc.org/ _______________________________________________ questions mailing list [email protected] https://lists.ntp.isc.org/mailman/listinfo/questions
