00:00:34 <jclulow> sommerfeld: Note that permissions on the socket do not affect the ability of users to connect to it
00:00:39 <jclulow> Even though sometimes it seems like they do
00:03:35 <jclulow> danmcd: That does seem like a good plan
00:04:15 <jclulow> Especially now that so many of the clocks we use are computerised, with NTP and up to date timezone databases, etc
00:10:39 <sommerfeld> jclulow: while making this work, I definitely got permission denied errors which ppriv -l attributed to a zfs_*access function.
00:11:04 <sommerfeld> and changing the permissions fixed the issue.
00:11:10 <jclulow> It depends on whether the software is using xpg4 or native sockets, unfortunately
00:11:30 <jclulow> (as to whether we bother to do that access check)
00:11:33 <jclulow> that is, the _client_
00:13:20 <sommerfeld> ah, right, I see that now.
00:13:33 <sommerfeld> I locked down the containing directory as well.
00:15:56 <jclulow> Yeah, that's what we recommend