Polling seems to have stopped, but only for my IPv4 address

Server operators
1

I have both an IPv4 and an IPv6 address. Long story short, I had to make some physical changes in my rack to accommodate a new device and had to shut everything down. Once I restored everything, I noticed that I stopped getting PPS signals, so I had to troubleshoot that which caused some additional time and instability in my data. Once that problem was resolved, the IPv6 score started to recover (pool.ntp.org: Statistics for 2600:1700:13ba:80f:27ad:9aca:96ec:b312). However, the IPv4 score has not moved at all (pool.ntp.org: Statistics for 104.176.153.167). I did verify that IPv4 is working and accurate from outside my network by putting the IPv4 address in a NTP checker on my phone and connecting over the cellular network. I am not sure what else can be done on my side.

2

Hi Nathan,

i can’t reach your IPv4 from any of my locations.

make sure ntp / chrony is running and listening on the public ip.
ex chrony:

$ ss -lpnu sport = :123
State           Recv-Q          Send-Q                   Local Address:Port                    Peer Address:Port
UNCONN          0               0                              0.0.0.0:123                          0.0.0.0:*              users:(("chronyd",pid=8267,fd=7))
UNCONN          0               0                                 [::]:123                             [::]:*              users:(("chronyd",pid=8267,fd=8))
3

I also get no response from the IPv4 from my hosts. When I traceroute IPv4 using port 123, I get ICMP TTL responses all the way to 104.176.153.167.

Run tcpdump/wireshark on the server and see if it is receiving any queries.

4 
 ss -lpnu sport = :123
State                  Recv-Q                 Send-Q                                 Local Address:Port                                 Peer Address:Port
UNCONN                 0                      0                                            0.0.0.0:123                                       0.0.0.0:*
UNCONN                 0                      0                                                  *:123                                             *:*

16:53:52.091907 IP 54.163.129.233.123 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.092311 IP 192.168.0.8.123 > 54.163.129.233.123: NTPv4, Server, length 48
16:53:52.208050 IP 75.177.214.46.123 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.208439 IP 192.168.0.8.123 > 75.177.214.46.123: NTPv4, Server, length 48
16:53:52.419449 IP 174.106.116.113.123 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.419854 IP 192.168.0.8.123 > 174.106.116.113.123: NTPv4, Server, length 48
16:53:52.422950 IP 173.70.84.94.123 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.423300 IP 192.168.0.8.123 > 173.70.84.94.123: NTPv4, Server, length 48
16:53:52.554912 IP 69.73.120.202.123 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.555458 IP 192.168.0.8.123 > 69.73.120.202.123: NTPv4, Server, length 48
16:53:52.767335 IP 54.208.194.128.123 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.767806 IP 192.168.0.8.123 > 54.208.194.128.123: NTPv4, Server, length 48
16:53:52.909087 IP 207.159.121.115.58263 > 192.168.0.8.123: NTPv4, Client, length 48
16:53:52.909662 IP 192.168.0.8.123 > 207.159.121.115.58263: NTPv4, Server, length 48

Thanks all. I think there was an iptables state that was in memory prior to shutdown but was not committed to the rules file. Adjusting the DROP rules in the IPv4 table fixed it. I confirmed it is saved now. :slight_smile: