NTP stopped working, unable to resolve host, time is wrong.

Hello! It seems that some update (perhaps within the past few months) has caused NTP to stop working.

I get errors like the following in my Arch Linux instance, as seen with journalctl -u ntpd.service:

Apr 11 22:41:03 hypership ntpd[17860]: giving up resolving host 1.arch.pool.ntp.org: Servname not supported for ai_socktype (-8)
Apr 11 22:41:03 hypership ntpd[17860]: giving up resolving host 2.arch.pool.ntp.org: Servname not supported for ai_socktype (-8)
Apr 11 22:41:03 hypership ntpd[17860]: giving up resolving host 3.arch.pool.ntp.org: Servname not supported for ai_socktype (-8)
Apr 11 22:41:05 hypership ntpd[17860]: giving up resolving host 0.arch.pool.ntp.org: Servname not supported for ai_socktype (-8)

but I have not changed anything. The only thing I've done is accept the standard Linode upgrade to the machine, but I have not modifies my Arch Linux instance.

It seems the time difference between my server and Google's is off by a couple minutes, which causes Google authentication to fail.

Has anything changed, like the way machines are booted? Are they chrooted/jailed/contained differently now? I'm asking because various search results, like http://opensuse.14.x6.nabble.com/ntpd-failing-td5060101.html , indicate problems when in a chroot.

Any ideas why those errors are happening?