I want my Slackware machine to be an NFS server so clients can connect to it. I have followed this http://docs.slackware.com/howtos:net...nd_dirty_setup and it does not work, I have also read this guide, which seems to be very outdated http://docs.slackware.com/howtos:net...home_nfs_howto and so I don't think is the correct way.

I have a few ports added on my firewall, but I went as far as to allow all incoming/open all ports on iptables to see if that would solve the issue but it didnt. I have made sure rc.nfsd and rc.rpc have executable bit and are running. My /etc/export file looks like this

I also typed exportfs -a and exportfs shows this directory being exported.

On various client machines I can not connect to the NFS server, if I try to connect to the NFS server on the same machine that is hosting it I get

I doubt I'll be able to help as I've only set up NFS a few times and it's been a while. It might help if you gave more details on what happens, if anything, when you try to connect. "It does not work" is too vague. Any error messages for example?

I'm using NFS in conjunction with NIS for centralized user management. Here's a little HOWTO I wrote on the subject for Slackware 14.0, which remains valid on 14.1. Just take the NFS bits and leave out the NIS stuff.

http://docs.slackware.com/howtos:net...aming_profiles

Cheers,

Niki

NFS is one of those things that I've run into very few problems with. I have quite a number of NFS exports and haven't touched iptables. Did you take anything from your CentOS install? On my machine, all I did was just make sure rc.nfsd and rc.rpc are both executable and then add my exports

I've mounted these as nfs and accessed them in xbmc/kodi through their nfs browser. Sure, it's insecure, but it's just on a local LAN at my house.

It might be worth making yours more insecure to see if it's a problem with permissions rather than nfs itself. Have you tried removing your 192.168.18.0/24 restriction (replace it with an asterik "*")? Also, try clearing out your /etc/hosts.allow and /etc/hosts.deny to make sure there is nothing preventing access. Once we know the server is working, we can then look at locking it down.

Well the thing is I don't really know how to give you an error message. I have tried mounting it via a client machine running arch linux via autofs (When my server was running CentOS my arch linux client mounted the NFS through autofs just fine) and it just gives me that it can't find such a directory. When I try through my Raspberry Pi running Debian 7 with Kodi I get the same similar message that it can't find the directory. I try through my phone which runs Android and Kodi and same problem. All of these clients were able to connect to my NFS the same way when the server was on CentOS.

I read this before, and it was very confusing for me as none of that is required in this HowTo http://docs.slackware.com/howtos:net...nd_dirty_setup
I will definitely open up those ports on the firewall though as I do not have all of those open (I have run with the firewall off and clients still couldnt connect to the nfs though)

Do I really need to assign a custom port for each of these? If so then I will do it, but it just seems strange to me that I have to.

I never had any problems with NFS on CentOS either, the thing always worked, occasionally you would have to restart it once in a while. The only thing I "took" from CentOS were the shares in /etc/exports as they were working great/fine when it was running CentOS. My hosts.allow and hosts.deny are and have been empty since install. I havent tried with just an asterisk, although all of my devices are on that subnet so it shouldn't be a problem really.


EDIT: Checked /var/log/syslog and I am getting this

What do find outdated in that guide? The information presented is still applicable to Slackware-14.1(stable) and Slackware-current.

I thought this one might be been older because I have to add so much more. I have to add things in hosts.allow and/or hosts.deny and I have to set a port for a bunch of services, however the quick and dirty guide does not require me to do that. So that led me to believe this guide was for an older Slackware release. Also because it talks about rpcinfo and how it is no longer available in 14.0

The additional configuration is about enhancing security and use of specific ports for use with a firewall for the reasons explained in the guide.
The rpcinfo command is deprecated, in favour of pmap_dump. Upstream disabled compilation of rpcinfo in the source used for Slackware 14+

TCP wrappers (hosts.allow and hosts.deny) are officially deprecated. So don't bother, just leave them empty.

Just as a matter of interest, have you considered using sshd (on the server - surely there already) and ssh-fuse (on the client)? It's far more secure than NFS, it's simpler to setup, and you get a finer control over user permissions. And it's pretty much stateless - I've seen NFS client machines do bad things when the NFS server "goes away" unexpectedly. Also look at autofs - a very clever, but slightly tricky thing to setup. The neat thing, even for NFS clients, is that when the client is no longer "on the share", it can be automagically disconnected. And then reconnected when needed - transparently.

A pragmatic question just out of curiosity. Things seemed to work well for you under CentOS, so why migrate to a distribution that's very different? Here, I'm using Slackware because it's what fits best, but then, YMMV.

I am not Altiris, but a partial answer is here.

@Altiris: notwithstanding what I wrote and that still stands, you seem to learn fast, congrats! After all everyone has one's own way of learning.

I have heard of these services very briefly, maybe I will look into in the future? What I need NFS for is just to be able to stream video files from my server to my Raspberry pi, NFS did the job fine and I don't need to give read/write permissions and its local also. I don't think I need sshd and these other stuff?

Yes, everything was working as it should under CentOS and I very much liked it and still do, however I think it was maybe a bit of wanting to try something new but I started messing with Slackware and I like the philosophy and approach to how things are maintained/kept in Slackware more than CentOS. Also, security updates were taking a really long time to get patched into CentOS. I remember in June there were 3-7 openssl vulnerabilities, I checked the security advisors on Slackware site and they had already issues patches, I only got my patches on CentOS about 6 days later. Theres also a big bug in CentOS due to libuser, which I don't believe Slackware uses? I understand why there is a delay for CentOS (first have to wait for rhel, then CentOS devs have to push it) but I just like Slackware more. I get sort of a warm feeling if my server is running Slackware than CentOS, I just like the philosophy behind Slackware more.