Brian Utterback wrote: > On 11/18/10 17:22, James Carlson wrote: > >> I had to do "ntpq -4p" to get it to work; the new tools seem to be funny >> about IPv6 loopback and my configuration. >> >> But the command did succeed, and printed out the peers as expected. It >> looks like the log file spam may have stopped. It's been 5 minutes >> since I restarted ntpd, and no new error messages are coming out. >> >> So why are names bad and numbers ok? >> > > Well, names have to go through the async resolver child process, with > a more complicated code path to get configured. Numbers just get used > directly. You and I have discussed the methods of network > configuration NTP uses in the past. The most recent community version > has gotten so complex in this area that I am very reluctant to ever > upgrade the version in Solaris to much beyond what it is now. It > should be very simple, but the desire to allow it to run on platforms > that do not support any modern network API's has blown the code way > out of proportion.
OK, understood -- I remember those discussions. But I'm still a little confused about what "external interfaces" might be and how they'd factor in. I have many interfaces on the system, two of which (one IPv4, the other IPv6) connect to "the Internet," so I'm pretty sure that if you're talking about IP interface lists, I've got the right ones configured in there somewhere. Could it be that ntpd is just starting "too early" and due to some internal defect, it fails to notice that interfaces eventually get plumbed? Could it be that just plain restarting ntpd after the dust settles is enough to shut this error message up? (I'll give that a try -- switching back to names rather than numbers and restarting ntpd.) For what it's worth, I'm using physical:default and not NWAM on this machine. Everything is statically configured via /etc/hostname* files. I haven't looked too closely yet at the SMF dependencies for NTP. I'm still recovering from the shock of having the upgrade completely blow away the cswsendmail service. :-/ -- James Carlson 42.703N 71.076W <carls...@workingcode.com> _______________________________________________ networking-discuss mailing list networking-discuss@opensolaris.org
