From Linux NFS

from CITI documentation.

NFSv4 configuration

Base configuration

This section maybe skiped on most recent distribution. If so, go here

You should have already built and installed the kernel and user utilities and setup krb5.

The following steps need only be done once. It maybe skiped most recent distribution.

Add the following lines to /etc/fstab to mount the rpc_pipefs and nfsd filesystems automatically:

rpc_pipefs	/var/lib/nfs/rpc_pipefs	rpc_pipefs	defaults	0	0
nfsd	/proc/fs/nfsd	nfsd	defaults	0	0

and create the mountpoint for rpc_pipefs, and mount both filesystems:

• mkdir /var/lib/nfs/rpc_pipefs
• mount rpc_pipefs
• mount nfsd

Modify the fstab line for filesystem you plan to export on the server to add the "acl" option to the mount options.

Create a directory to export on the server:

• mkdir /export
• chmod a+rwxt /export

The following steps need to be performed on every boot; this should really be done by init scripts. The necessary init scripts already exist in recent Fedora distributions. See also debian/nfs-common.init and debian/nfs-kernel-server.init in the patched nfs-utils tarball for example init scripts.

On your server, run rpc.mountd, rpc.svcgssd, rpc.idmapd, and rpc.nfsd:

rpc.mountd
rpc.idmapd
rpc.svcgssd   -> this is obsolete as of nfs-utils version 1.0.8-1
rpc.nfsd 8

On the client, run rpc.gssd and rpc.idmapd:

rpc.gssd -m
rpc.idmapd

- - - - - - - - -

Here are some distribution specific guidelines which have been tested on openSUSE and Ubuntu. They attempt to clarify the installation of secure NFS4 on a Linux systems.

Source: Steve Sheriff, steve.lcb@gmail.com

Please note that I am not a Linux expert and that I do not usually use English in my day to day life.

These notes were written after I faced the task of installing secure NFS4 for a heterogeneous LAN of openSUSE, Ubuntu and Windows clients under a Samba 4 Active Directory PDC. This was tested and reproduced on budget hardware. The largest cost for any single component of the LAN was the licence to activate Windows 7 Professional, needed to join Microsoft clients to AD.

I have written this in an attempt to encourage all of us who find ourselves in similar administrative roles, to maintain a centralised deposit of correct information, written in plain English, on modern Linux distributions. I also hope that it will serve to dispel many of the myths I read during the installation and save the countless forays I made into uncharted territory.

Myth: NFS4 must be exported from an fsid=0 pseudo root directory. e.g. I am exporting the /home directory to the clients. The /etc/exports file could be either of these examples.

e.g. 1. How not to do it /home is bind mounted to chmod'ed 1777 /exports /etc/exports /export gss/krb5(rw,fsid=0,insecure,no_subtree_check,crossmnt) /export/home gss/krb5(rw,insecure,no_subtree_check) /export gss/krb5i(rw,fsid=0,insecure,no_subtree_check,crossmnt) /export/home gss/krb5i(rw,insecure,no_subtree_check) /export gss/krb5(rw,fsid=0,insecure,no_subtree_check,crossmnt) /export/home gss/krb5(rw,insecure,no_subtree_check) /export *(rw,fsid=0,insecure,no_subtree_check) /export *(rw,nohide,insecure,no_subtree_check)

e.g. 2 . Simply use the familiar nfs3 export method. /etc/exports /home *(rw,sec=none:sys:krb5:krb5i:krb5p,no_subtree_check)

Choice of 5 flavours for mounting on the client.

Please see the comment above: 'The linux implementation allows you. . .'

Myth: you must use the nohide option for the bind mount to be visible on the client. You can use crossmnt on the pseudoroot or nohide on the bind mount. Or neither and use simply the conventional nfs3 syntex for /etc/exports.

Myth: NFS4 can only use weak DES cryptography e.g. you have been told that NFS4 can only use weak DES cryptography and so you need the line: [libdefaults] allow_weak_crypto = true added to /etc/krb5.conf Not needed. Tested using arcfour-hmac-md5 with kernels 3.1.9-1.4-desktop (openSUSE 12.1) and 3.0.0-15-default** (Ubuntu 11.10) **need to check the uname -r output on Ubuntu. Am writing his on openSUSE

Myth: NFS4 needs only the nfs/fqdn@REALM principal in it's keytab No. It needs the machine principal too. rpc.gssd will not allow you to make a test mount on the server (e.g. to test by mounting to /mnt) unless the machine principal is there along with the nfs principal in the keytab.

Myth: The client must have a nfs/client.domain principal in its No. It doesn't need to have one. From man rpc.gssd(8) Previous versions of rpc.gssd used only "nfs/*" keys found within the keytab. To be more consistent with other implementations, we now look for specific keytab entries. The search order for keytabs to be used for "machine credentials" is now: <HOSTNAME>$@<REALM> root/<hostname>@<REALM> nfs/<hostname>@<REALM> host/<hostname>@<REALM> root/<anyname>@<REALM> nfs/<anyname>@<REALM> host/<anyname>@<REALM>

A real life example. hh3.hh3.site is my server and HH6$@HH3.SITE is the machine principal for the client. On the client, issue a: mount -t nfs4 hh3:/foo /bar -o sec=krb5

Kerberos: ENC-TS Pre-authentication succeeded -- HH6$@HH3.SITE using arcfour-hmac-md5 Kerberos: AS-REQ authtime: 2012-02-06T19:44:47 starttime: unset endtime: 2012-02-07T05:44:47 renew till: 2012-02-07T19:44:47 Kerberos: Client supported enctypes: aes256-cts-hmac-sha1-96, aes128-cts-hmac-sha1-96, des3-cbc-sha1, arcfour-hmac-md5, des-cbc-crc, des-cbc-md5, des-cbc-md4, using arcfour-hmac-md5/arcfour-hmac-md5 Kerberos: Requested flags: renewable-ok Kerberos: TGS-REQ HH6$@HH3.SITE from ipv4:192.168.1.10:45421 for nfs/hh3.hh3.site@HH3.SITE [canonicalize, renewable] Kerberos: TGS-REQ authtime: 2012-02-06T19:44:47 starttime: 2012-02-06T19:44:47 endtime: 2012-02-07T05:44:47 renew till: 20

But remember from above, that the nfs server must be able to identify itself. In this example, klist -ke /etc/krb5.keytab Keytab name: WRFILE:/etc/krb5.keytab KVNO Principal

---

---------

  1 nfs/hh3.hh3.site@HH3.SITE (arcfour-hmac)  
  1 HH3$@hh3.site (arcfour-hmac)

rpc processes 1. openSUSE after starting the nfs server

ps aux | grep rpc root 1231 0.0 0.1 2356 764 ? Ss 08:20 0:00 /sbin/rpcbind root 3737 0.0 0.0 0 0 ? S< 08:28 0:00 [rpciod] root 5215 0.0 0.2 3684 1208 ? Ss 09:26 0:00 /usr/sbin/rpc.svcgssd root 5223 0.0 0.0 2416 344 ? Ss 09:26 0:00 /usr/sbin/rpc.idmapd root 5227 0.0 0.1 2976 748 ? Ss 09:26 0:00 /usr/sbin/rpc.mountd --no-nfs-version 2 --no-nfs-version 3 (note the lack of gssd)

2. On both openSUSE (after calling rpc.gssd) and Ubuntu with all the necessary processes for successfully exporting the directory with -o sec=krb5<flavour>

ps aux | grep rpc root 1231 0.0 0.1 2356 764 ? Ss 08:20 0:00 /sbin/rpcbind root 3737 0.0 0.0 0 0 ? S< 08:28 0:00 [rpciod] root 5215 0.0 0.2 3684 1208 ? Ss 09:26 0:00 /usr/sbin/rpc.svcgssd root 5223 0.0 0.0 2416 344 ? Ss 09:26 0:00 /usr/sbin/rpc.idmapd root 5227 0.0 0.1 2976 748 ? Ss 09:26 0:00 /usr/sbin/rpc.mountd --no-nfs-version 2 --no-nfs-version 3 root 5294 0.0 0.0 3816 428 ? Ss 09:31 0:00 rpc.gssd

Installation details 1. openSUSE Packages: zypper in nfs-kernel-server nfs-client nfsidmap limal-nfs-server

Configuration (Or use the excellent Yast NFS-Server Module) /etc/sysconfig/nfs USE_KERNEL_NFSD_NUMBER="4" MOUNTD_PORT="" NFS_SECURITY_GSS="yes" NFS3_SERVER_SUPPORT="no" NFS4_SUPPORT="yes" SM_NOTIFY_OPTIONS="" NFS_START_SERVICES="yes" STATD_OPTIONS="" NFSV4LEASETIME="" RPC_PIPEFS_DIR="" SVCGSSD_OPTIONS="" NFSD_OPTIONS=""

2. Ubuntu Packages: apt-get install nfs-common nfs-kernel-server

Configuration /etc/default/nfs-common NEED_STATD= STATDOPTS= NEED_IDMAPD=yes NEED_GSSD=yes

/etc/default/nfs-kernel-server RPCNFSDCOUNT=8 RPCNFSDPRIORITY=0 RPCMOUNTDOPTS=--manage-gids NEED_SVCGSSD=yes RPCSVCGSSDOPTS= RPCNFSDOPTS=

Needed on both client and server on both openSUSE and Ubuntu: /etc/idmapd.conf [General] Verbosity = 0 Pipefs-Directory = /var/lib/nfs/rpc_pipefs Domain = hh3.site [Mapping] Nobody-User = nobody Nobody-Group = nogroup

Debugging Run rpc.idmapd -fvvv and rpc.gssd -fvvv in separate terminals. They will tell you in no uncertain terms when you've got it wrong.

rfc It would be helpful if we could gather together our experiences on NFS4 installation and centralise them here. Particularly helpful would be distribution specific contributions with working examples. My suggestion would be one page per distribution.

References: 1. This wiki:

http://wiki.linux-nfs.org/wiki/index.php/Nfsv4_configuration

2. man rpc.gssd(8)

http://linux.die.net/man/8/rpc.gssd

3. A hands on practical using modest hardware. Includes screenshots. openSUSE specific.

http://linuxcostablanca.blogspot.com/p/samba-4.html

4. The same as (3) but with Ubuntu specific diffences

http://linuxcostablanca.blogspot.com/2012/01/samba-4-ubuntu.html

5. The latest unofficial version of this wiki page. Any new findings, I'll post there.

http://linuxcostablanca.blogspot.com/2012/02/nfsv4-myths-and-legends.html

Exporting directories

Export your server directory (in this case, to anyone, using any authentication flavor), by adding the following lines to /etc/exports

/export  gss/krb5(rw,fsid=0,insecure,no_subtree_check)
/export  gss/krb5i(rw,fsid=0,insecure,no_subtree_check)
/export  gss/krb5p(rw,fsid=0,insecure,no_subtree_check)

and then running

exportfs -r

(Note that if your server is connected to the internet the first line will give everyone in the world read and write access to /export and (with a little work) to anything on the same disk partition as /export! See the exports man page for more information.)

To access this export from a client:

Create a mountpoint on your client:

mkdir /mnt/nfs4/

and mount the server:

mount -t nfs4 yourserver:/export /mnt/nfs4/

or, to mount with kerberos,

mount -t nfs4 -o sec=krb5 yourserver:/export /mnt/nfs4/

Play around with the mounted filesystem:

cd /mnt/nfs4/
echo "Hello World!" >FOO
cat FOO
getfacl FOO

etc.

To mount a filesystem using krb5, provide the "-osec=krb5" option to mount. To export a filesystem using krb5, export it to the special client named "gss/krb5". Replace "krb5" by "krb5i" or "krb5p" for integrity or for privacy, respectively. The -osec= options should also work for NFSv3 exports and mounts.

Exporting and mounting can be automated using /etc/fstab and /etc/exports; see the man pages fstab(5) and exports(5).

Tecnical note: NFSv4 no longer has a separate "mount" protocol. Instead of exporting a number of distinct exports, an NFSv4 client sees the NFSv4 server's exports as existing inside a single filesystem, called the nfsv4 "pseudofilesystem".

The linux implementation allows you to designate a real filesystem as the pseudofilesystem, identifying that export with the fsid=0 option; we no longer recommend this. Instead, on any recent linux distribution, just list exports in /etc/exports exactly as you would for NFSv2 or NFSv3.