[ntp:hackers] Further confirmation multicast broken

mayer at gis.net mayer at gis.net
Sun Feb 27 13:04:43 PST 2005


Dave,

I do have an idea that I'm testing out. We'll see shortly.

Danny
----- Original Message Follows -----
> Guys,
> 
> Multicast doesn't work at all. Confirmation follows.
> 
> In principle, multicast with symmetric key should work. However,
> because  of the bug reported previously, the volley to calibrate the
> propagation  delay fails because the multicast address, rather than
> the unicast  address, is in dstadr in the association. See the
> behavior on howland,  which is receiving multicast from two servers.
> The first time the  multicast is received, the association is
> mobilized correctly. It sends  a client message to the server as
> expected. Somewhere between there and  the server response the dstadr
> is changed to the multicast address and  the volley fails.
> 
> I repeat. The dstadr is set correctly in newpeer(). It is not changed 
> anywhere else in ntp_peer.c or ntp_proto.c. Something, either in a
> macro  or in ntp_io.c is changing it. It should never be changed by
> anything  after newpeer().
> 
> In principle, it would be possible to mobilize and continue the 
> assocation by including a novolley keyword on the multicast command in
> the configuration file. Unfortunately, the syntax of that command
> can't  currently do that. The items following the multicast keyword
> must be  only group addresses.
> 
> Dave
> _______________________________________________
> hackers mailing list
> hackers at support.ntp.org
> https://support.ntp.org/mailman/listinfo/hackers



More information about the hackers mailing list