-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > > > Shouldn't it be sending a KILL upstream? > > > > This means that the first server thats clock is slow will squit every > > other server on the network. Should be fun! > > Actually, an addendum to my last email--I couldn't decide whether to > put in cptr or cli_user(sptr)->server. Putting in cptr will squit *us* > from the network (if we're a leaf), so that might be better--it'll force > us to reconnect and thus resync.
Ah, cptr would be good. > I think I'd still prefer a solution where we solicit a SETTIME...or > better yet, make certain everyone's using ntpd and force them all to > turn on RELIABLE_CLOCK :) (At this rate, I tend to prefer the latter; > we're having far, FAR too many clock problems :/ ) Hmm, sounds like we should send a "SETTIME", then if we ever recieve a SETTIME thats wrong, and our clock is RELIABLE, then we send a SETTIME back upstream with the correct time instead. Unless we have wrong RELIABLE_CLOCK's this would work nicely. If the clocks aren't reliable, then they could "war" - -- The worst cliques are those which consist of one man. -- G.B. Shaw -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) Comment: Only when you are sure they have you, can you stop being paranoid iD8DBQE94UdlcAgRpy8z8UQRAibIAKC71tOiy5ZcqG45cvXry0TOZ6+bfwCg0cou VgJCki8W23GNXeuBoiaLxC4= =UDX0 -----END PGP SIGNATURE-----