I just ran some stats on my S1 and S2 public & pool time servers.  The
results are interesting if a little scary.

This is in no way a statistically valid study - just curiosity on my part.

Dump the mon list from both servers (1200 hosts)
Eliminate all those that don't use port 123, internal hosts and servers
configured into my setup  (down to 603)
run ntpdate -q against them - 90 respond.
run ntpq -p against them - less than 30 respond.

Of those that respond they fall into two broad groups.
1) hosts with 6 to 8 stratum 1 or 2 servers configured (probably not
using the pool) - none of these were in the pool.
2) hosts with only 1 server configured (ugh!)

The remainder were oddballs which claimed not to have my server in their
config and three (yes three) hosts that had three or four servers which
looked like they came from the pool and were sanely configured.

I also ran 600 of the hosts against the pool stats server - the only hit
was on my own server I put in the list as a check.

These stats observations are clearly skewed by the fact that the well
configured machines are also likely to be behind a firewall..

I'm not drawing any major conclusions from this other than thinking I
may (if I get the time) write a tool to auto survey ntp clients that use
my servers.

John
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to