@spockfish Not sure where you prefer bug reports, so leaving this here.
The current ntp configuration on RoPieee is not ideal. It uses the “server” statement with pool.ntp.org. The problem with this, is ntp will solicit the pool members to use once at startup and NEVER AGAIN. It is quite common for ntp pool members to go away, leaving one trying to query dead IPs.
The proper way to address this is replace the “server” keyword with “pool”. (It is also generally considered best practice to include the “iburst” keyword at the end of each server statement.) Using the “pool” keyword, ntp will automatically replace time sources that have been unreachable for an extended period of time (something like 24 hours, I forget exactly). Note: for the “pool” keyword to work, however, you must also duplicate the existing “restrict default” line, change “default” to “source” and remove the “nopeer” keyword. (Do not delete the existing “restrict default” line.)
I made these changes to the RoPieeeXL I just set up for the first time this week. However, upon reboot, it appears your configuration script re-writes the ntp.conf file so it is back to using the suboptimal configuration.
Please consider updating your ntp template to switch to using the “peer” keyword.