NTP on AWS

- [2 mins read]

Ran into an issue where a Linux instance running on AWS in a private subnet was not updating the system time via NTP. First check was for the config file, but it had a list of servers, both inside and 1 outside AWS:

server 0.amazon.pool.ntp.org
server 0.us.pool.ntp.org
server 1.amazon.pool.ntp.org
server 2.amazon.pool.ntp.org

From this post (I would like to link to it, but it has been 8.5 years since I wrote this, and only found the missing link today on 2024/05/29) I tried both ntpdate and ntpdate-debian with the following results:

$ ntpdate
16 Dec 22:24:14 ntpdate[32112]: no servers can be used, exiting

$ ntpdate-debian
16 Dec 22:24:56 ntpdate[32117]: no servers can be used, exiting

The instance did have access to a NAT box to allow access outside of the VPC and ntpq -p returned the following:

$ ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 ranger.innolan. .INIT.          16 u    -   64    0    0.000    0.000   0.000
 4.53.160.75     .INIT.          16 u    -   64    0    0.000    0.000   0.000
 origin.towfowi. .INIT.          16 u    -   64    0    0.000    0.000   0.000
 ntp.your.org    .INIT.          16 u    -   64    0    0.000    0.000   0.000

Turns out that you need to explicitly allow udp/123 on the NAT instance’s Security Group. Once I added the rule, everything started working:

$ ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 216.152.240.220 216.218.254.202  2 u    1   64    1   37.486   11.927   0.138
 ranger.innolan. 90.184.222.115   3 u    2   64    1   73.323    9.543   0.025
 srcf-ntp.stanfo .shm0.           1 u    1   64    1   25.009   12.412   0.019
 grom.polpo.org  127.67.113.92    2 u    2   64    1   22.254   13.909   0.000

comments powered by Disqus