An ntpmon proof-of-concept exists
Eric S. Raymond
esr at thyrsus.com
Fri Nov 25 15:36:19 UTC 2016
As I expected, a first cut at ntpmon took only a couple of hours to
write. This is where we start to collect some serious benefits from
having a common back-end library for Python clients; these will
continue when we start on ntpshark.
A couple of problems stand in the way of landing ntpmon in the tree,
however. One is purely a housekeeping issue; rather than create a
ntpmon/ subdirectory that will go away in the upcoming directory
reorganization, I'd rather land Python ntpdig first, do the
reorg, and then commit ntpmon.
(Am currently thinking the new directory will be named "clients".)
The other problem is a bit more serious. I can't seem to get my ntpd
to report mrulist entries. This hampers testing of ntpmon; the lower
window is supposed to be an mrulist display. Right now I can only
demonstrate the upper half, the automatically-refreshing peers display.
To answer the obvious questions: No, I do not have nonrulist
configured, and my daemon thinks mru_enabled is 1 when I query it
through ntpq. And I can (sometimes) see MRUlist entries from Hal Murray's
instance at 22.214.171.124, so ntpd is generating them.
Hal thinks some configuration switch is implicated, but doesn't know
which. I don't either. Can anyone alse shed any light on this?
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
I have never made but one prayer to God, a very short one: "O Lord,
make my enemies ridiculous." And God granted it.
More information about the devel