[07:27] <JonathanD> Morning!
[09:20] <rmg51> Morning
[12:29] <teddy-dbear> Morning peoples, critters and everything else
[13:33] <WorkingTurkey> mornin
[19:01] <ChinnoDog> Anyone know a way to intentionally make NFS report less free space than it would otherwise?
[19:08] <ChinnoDog> hmm. I can use quotas but that looks complicated for such a simple problem.
[19:10] <jthan> ChinnoDog: What filesystem is the actual data on?
[19:11] <ChinnoDog> btrfs but that is only because it is my system. I can't assume it is any particular file system because it is part of a vagrant project.
[19:11] <jthan> Well... nfs does not actually "report" the available space
[19:13] <r00t^2> i think the base problem is you have these edge cases that you expect to be covered- why on earth would you want an NFS share to seem smaller than it is?
[19:14] <r00t^2> if it's because you don't want people to use more than X amount, then yes- that's exactly what quotas are for
[19:15] <ChinnoDog> No, it is a band-aid for a script that selects a directory to store its data.
[19:15] <ChinnoDog> It selects the volume with the largest free space by default but that volume is the NFS mount and causes the software to break because it can't put its data there.
[19:16] <r00t^2> bandaids aren't good. add exclusion/priority support in the script
[19:16] <ChinnoDog> I can't. It isn't my script.
[19:16] <ChinnoDog> Also, it is subject to updates so I can't modify it.
[19:17] <jthan> pin it at the current version? Or don't allow the script / program calling it to even know your nfs share exists
[19:17] <ChinnoDog> It is definitely a script shortcoming but nothing I can do anything about.
[19:17] <r00t^2> then file a bug with the maintainer of the script
[19:17] <ChinnoDog> Is there an easy way to deny permission to it to only one user?
[19:17] <r00t^2> oh, sure
[19:18] <ChinnoDog> Yes, that is a long term solution but not one that I can use right now.
[19:18] <r00t^2> https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Storage_Administration_Guide/ch-acls.html
[19:19] <r00t^2> or you can do it via standard octals and ownerships
[19:21] <ChinnoDog> I can see how I can use additive permissions for it but not deny permissions
[19:23] <jthan> make sure user:user is not a match to that using script
[19:23] <jthan> and then set appropriate octals
[19:24] <ChinnoDog> I need other users to be able to access it though. I'm not sure which ones. I will check and see if I can use additive permissions for just those users.
[19:24] <jthan> then make sure they are in :group
[19:25] <jthan> Folder "ubuntu" has ownership chinno:pa
[19:25] <jthan> jthan and Chinno are in group pa
[19:25] <jthan> bts is not
[19:26] <ChinnoDog> I think you are missing what I am trying to say but it doesn't matter. I should be able to identify which users need access.
[19:28] <jthan> You lost me, Chinno.
[19:32] <ChinnoDog> I wanted to deny permission to a single user which linux/unix doesn't support. It only supports an additive permissions model.
[19:35] <r00t^2> jthan: i was wrong, more than 5 seconds. btw.
[19:36] <jthan> ChinnoDog: I think you can do it with acls
[19:37] <ChinnoDog> I have never seen a subtractive linux ACL.
[19:39] <r00t^2> setfacl -m user:USERNAME:---
[19:39] <r00t^2> setfacl -m user:USERNAME:--- <file/directory/whatever>
[19:39] <r00t^2> ~*ta-da*~
[19:40] <r00t^2> https://wiki.archlinux.org/index.php/Access_Control_Lists
[19:41] <r00t^2> while you're at it, you should probably grab a copy of http://www.amazon.com/UNIX-Linux-System-Administration-Handbook/dp/0131480057/ref=sr_1_1?ie=UTF8&qid=1439494862&sr=8-1&keywords=unix+and+linux+system+administration+handbook
[19:47] <ChinnoDog> oh. The ACL is on top of the unix permissions. I didn't realize they weren't the same thing.
[19:47] <ChinnoDog> That is more complexity than I need. I think I can solve it without it.
[19:47] <jthan> Oh yeah - has to be installed but.. no real configuration and REALLY easy to use thereafter.
[19:49] <r00t^2> jthan: it's actually usually installed by default
[19:49] <r00t^2> but the fs needs to be *remounted* with support for them
[19:49] <r00t^2> or enabled via tune2fs
[19:50] <r00t^2> but yeah i mean.. you aren't really going to find something more simple than that.
[20:26] <ChinnoDog> No umask on nfs :-(