[lug] Fedora 19, "storage administrators guide", NFS

[Maxwell Spangler]

On Sat, 2013-11-30 at 01:45 +0000, stimits at comcast.net wrote:

> I'm trying to figure out why a fedora 19 system is failing to allow
> mount to another machine on a trusted private LAN.
> 
> My research finds that the most recent System Administrator's Guide at
> http://docs.fedoraproject.org is for fedora 18, although fedora 19 has
> been around a while and fedora 20 is about to go production. In this
> fedora 18 guide I find a reference suggesting NFS setup is covered in
> the Storage Administrator's Guide, which in turn is most recently
> updated to fedora 14. Much of the fedora 18 System Administrator's
> Guide is no longer applicable, and refers to non-systemd
> configuration.
> 
> Has anyone here found any kind of NFS setup guide which applies to the
> systemd (systemctl) method of configuration on fedora 19 or newer? Is
> there any kind of check list available for NFS v4 on a systemd setup
> such as seen on fedora 19?
> 
> I have put the private LAN on the "trusted" zone, and all other
> networking succeeds here, but I get no log messages to indicate any
> attempt to mount my exports and the system doing the mount simply
> times out...hoping to find clues or a place to start which uses the
> actual fedora 19 packages which are not antiques. Fedora 14 and 18
> documents seem to no longer apply.

I run into this problem from time to time.  Given the complexity of a
Linux operating environment these days you need to be patient, approach
the problem slowly and dig down into the layers of what allows it to
work.

You left out a key component: What is the other system? How is it
configured?

If you have a firewall running on the server, this can get in the way of
NFS1-3 due to the way ports are handled.  Disable the firewall for now,
to ensure that's not the problem.

Go to your server and verify that nfs is running


➜ root at filer [~] # ps axf | grep nfs
 5170 ?        S      0:00  \_ [nfsd4]
 5171 ?        S      0:00  \_ [nfsd4_callbacks]
 5172 ?        S      0:15  \_ [nfsd]
 5173 ?        S      0:14  \_ [nfsd]
 5174 ?        S      0:13  \_ [nfsd]
 5175 ?        S      0:15  \_ [nfsd]
 5176 ?        S      0:16  \_ [nfsd]
 5177 ?        S      0:15  \_ [nfsd]
 5178 ?        S      0:14  \_ [nfsd]
 5179 ?        S      0:17  \_ [nfsd]
 7511 ?        S      0:00  \_ [nfsiod]
 7545 ?        S      0:00  \_ [nfsv4.0-svc]
 8479 pts/0    S+     0:00                      \_ grep nfs


You can also check: (This is Centos/Red Hat 6.4)


➜ root at filer [~] # service nfs status
rpc.svcgssd is stopped
rpc.mountd (pid 5116) is running...
nfsd (pid 5179 5178 5177 5176 5175 5174 5173 5172) is running...
rpc.rquotad (pid 5112) is running...


Check your /etc/exports to ensure that they allow access to your client.
This example offers access to an HP Elite model desktop named, 'elite':


/srv/filer_general elite(rw,async,no_root_squash)


On the client, see if you can see the server:


➜ maxwell at elite [~] $ showmount -e filer
Export list for filer:
/srv/filer_general     elite.fossil


Try mounting it:


➜ root at elite [~] # mount -t nfs filer:/srv/filer_general/iso /mnt/iso


If you find something interesting in your environment (like errors,
etc), post it so we can comment.

-- 
Maxwell Spangler
========================================================================
Linux & Open Source Systems Engineer
Fort Collins, Colorado, USA
http://www.maxwellspangler.com/


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lug.boulder.co.us/pipermail/lug/attachments/20131129/1e70b47e/attachment.html>