Revert "nixos-container: use systemd-run instead of nsenter"

This reverts commit 7cb100b.

This appears to break at least the `container`-backend of `nixops`: when
running `switch-to-configuration` within `nixos-container run`, the
running `systemd`-instance gets reloaded which appears to kill the
`systemd-run` command and causes `nixos-container run` to hang.

The full issue is reported in the original PR[1].

[1] NixOS#67332 (comment)

pkgs/tools/virtualization/nixos-container/nixos-container.pl

@@ -9,6 +9,7 @@
 use Cwd 'abs_path';
 use Time::HiRes;
 
+my $nsenter = "@utillinux@/bin/nsenter";
 my $su = "@su@";
 
 # Ensure a consistent umask.
@@ -319,10 +320,9 @@ sub restartContainer {
 # Run a command in the container.
 sub runInContainer {
     my @args = @_;
-
-    exec("systemd-run", "--machine", $containerName, "--pty", "--quiet", "--", @args);
-
-    die "cannot run ‘systemd-run’: $!\n";
+    my $leader = getLeader;
+    exec($nsenter, "-t", $leader, "-m", "-u", "-i", "-n", "-p", "--", @args);
+    die "cannot run ‘nsenter’: $!\n";
 }
 
 # Remove a directory while recursively unmounting all mounted filesystems within