Usually when you have a non default DNS configuration in your system, for example if you're using dnsmasq or another DNS service, other than systemd-resolve, it's possible that dirmngr used by gpg fails to get the resolved name for, then, you need to check your name resolution software. # See man:rvice(8) for details about the supported modes of # different way, replace this symlink by a static file or a different symlink. # Third party programs should typically not access this file directly, but only # Run 'resolvectl status' to see details about the uplink DNS servers # internal DNS stub resolver of systemd-resolved. # This is a dynamic nf file for connecting local clients to the

# This file is managed by man:systemd-resolved(8). Here are my recommended commands and output to make the symbolic link: $ # you should check this has reasonable contents before using it I suggest reading man 8 systemd-resolved, section '/ETC/RESOLV.CONF' for details on what this does, exactly. You want to make a symbolic link to systemd's autogenerated nf that reflects the current resolver settings. etc/nf has something like hosts: files mymachines myhostname resolve dns in it, where resolve is a module provided by systemd (see man 8 libnss_resolve.so.2 for details). Note that dig, drill, nslookup all read nf directly, so these may not work. etc/nf is empty but hostname resolution e.g. Gpg: keyserver receive failed: Server indicated a failure I assume that it's a case of gpg not using nsswitch and the associated libc functions and instead, for some reason, reading nf itself.

This problem can be caused (as I just experienced it) by an empty nf, as may be the case on a system using systemd-resolved for its primary DNS resolution via nsswitch.

Zimbra install sh permission denied