[ntp:questions] Problem with NTP multicast server configuration

Larry Mills lgmills at claritycsi.com
Sat Jan 15 22:47:58 UTC 2005


I am having trouble establishing a reliable NTP multicast server
configuration.

First off, let me describe the configuration I am trying to achieve:

- All hosts to be time synchronized are on a closed network; absolute time
synchronization with a true stratum 1 source is not important, but all
servers
on the network should be closely synchronized with each other.

- Two hosts (hosts X and Y) are configured as NTP multicast servers; each
server is a peer of the other, in symmetric-active mode.

- All other servers are configured as multicast clients; the only keyword in
the client's /etc/inet/ntp.conf file is "multicastclient".

- The desire is for each server to be able to generate NTP (multicast)
broadcasts and respond to multicast client queries in cases where the peer
server is unreachable/unavailable.

- In cases where the peer server is unavailable, the other peer will revert
to its internal clock as the reference time.

The ntp.conf file for the simple configuration I have set up is below.  The
configuration file is identical in both peer servers, with the exception
that
the "peer" line contains the IP address of the other server.

####################### ntp.conf #############################

peer (address of other NTP server)

server 127.127.1.0

fudge 127.127.1.0 stratum 10

broadcast 224.0.1.1 ttl 4

driftfile /var/ntp/ntp.drift

####################### end ntp.conf ##########################

The problem I am seeing is that when only one server is available, for
example,
if xntpd is started on host X, and host Y is not up, no multicast broadcasts
are sent from the server, and multicast queries (e.g. from ntpdate) from
clients are ignored.  While in this state, ntpq -p shows that the server
is using it's local clock, and the peer information is discarded.

Once the second server (Y) becomes available and exchanges timing
information
with the peer, the second server (only) begins to issue multicast
broadcasts.
Output from ntpq -p shows that the second server is using the first as its
timing reference.  Even after the second server begins to issue NTP
multicasts
the first server continues not to issue them.   This behavior is symmetric, 
that is, if server Y is the first active NTP server, it will not
issues multicasts, and server X will issue them as soon as it synchronizes
with
Y.

All hosts in the cluster of servers are SPARC-based, running Solaris 9.

I'm certainly not a NTP expert, but I think the configuration I'm trying
to implement is possible, so I'm assuming I've botched something in the
configuration file.  Any help on this problem would be greatly appreciated!

Larry



More information about the questions mailing list