cluster:Nfs/3

From Dgiref

Please open a NGI-DE ticket if you experience any Installation or Configuration problem.

Nfs/3/server

Prepare

Operating system: Scientific Linux version 4.7 64 bit

Optimizing the configuration: 

Use minimal operating system installation without firewall.
To verify installed packages use the command 
rpm -qa | grep package_name
Install the following additional packages:
yum -y install wget yum rpm make gcc gcc-c++ tar sed zlib openssl
After the installation is complete, turn off any unnecessary services (like gpm, sendmail, cups, haldaemon, messagebus, pcmcia, anacron, atd) with the following command: 
chkconfig <SERVICE> off

Configure the following settings for the server:

Prepare new hard disk

To export a brand new disk through nfs, the corresponding device in the /dev directory must be find (use the boot log messages from /var/log). Lets assume that the new device is /dev/hdg.
Create the new partition on your disk with the command: fdisk /dev/hdg. Follow the instructions and create one big partition /dev/hdg1.
Format the new partition with the ext3 file system with the following command: mkfs -t ext3 -c /dev/hdg1 ("-c" options search for invalid disk sectors).
Create the new export directory (/srv/nfs) and change the permission to 1777 and mount the partition on it.

The new device will be mounted on boot on the selected directory.

Firewall configuration

configure TCP port 2049 (how to open port in firewall).

administrator's script: prepare.sh

```
 
```
```
#!/bin/bash
```
```
# prepare the nfs server installation
```
```
# execute as root
```
```
# echo `su -`
```
```
 
```

# Declare the variables section ------------

# Please insert your actual configuration

# NFS_SERVER_DIR="export directory on nfs server"

# HD_DEVICE="export device on nfs server"

# RIGHTS="access mode for export directory on the nfs server"

```
# Private network adresses
```
```
# NETWORK=ip adress for private network
```
```
# NETWORK_MASK=private network mask
```

# from here ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

```
NFS_SERVER_DIR="/srv/nfs"
```
```
HD_DEVICE="/dev/sdb1"
```
```
RIGHTS="1777"
```
```
NETWORK=10.0.171.0
```
```
NETWORK_MASK=255.255.255.0
```

# till here ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

```
 
```
```
#-> start routine
```
```
mkdir -p ${NFS_SERVER_DIR}
```
```
chmod ${RIGHTS} ${NFS_SERVER_DIR}
```
```
#<- end routine
```

Install

Probably, the operating system has nfs service installed. To be sure, the following packages will be installed:

nfs-utils
portmap
quota

administrator's script: install.sh

```
 
```
```
#!/bin/bash
```
```
# install nfs server
```
```
# execute as root
```
```
# echo `su -`
```
```
# load parameters from prepare section
```
```
cd `dirname $0`
```
```
source prepare.sh
```
```
 
```
```
#-> start routine
```

# Verify that the following packets are not installed with the command:

```
echo `rpm -qa | grep nfs-utils` 
```
```
echo `rpm -qa | grep portmap`
```
```
echo `rpm -qa | grep quota`
```
```
 
```
```
yum -y install nfs-utils portmap quota
```
```
#<- end routine
```

Configure

The NFS server has to be configured to export filesets, like the users home directories, the system software (etc/profile.d, ogsadai), the VO specific software areas, the UNICORE filespace. Server in D-Grid reference installation exports only one directory, located into /srv/nfs with the following attributes:

	Export	Export options	Node	Mountpoint	Mount options
nfs	/srv/nfs	rw,async	alle	/srv/nfs	nosuid

Another requirements are to configure /etc/hosts.deny and /etc/hosts.allow configurations files. Add lines like the following (template: daemon_name: ALL):

cat /etc/hosts.deny
portmap: ALL
lockd: ALL
statd: ALL
mountd: ALL
rquotad: ALL

In the file /etc/hosts.allow the lines for each of the five daemons:

portmap
lockd
statd
mountd
rquotad

should look like this (template: daemon_name: ip.pc.client1, ip.pc.client2):

cat /etc/hosts.allow
portmap: ip.pc.client1, ip.pc.client2
lockd: ip.pc.client1, ip.pc.client2
statd: ip.pc.client1, ip.pc.client2
mountd: ip.pc.client1, ip.pc.client2
rquotad: ip.pc.client1, ip.pc.client2

administrator's script: configure.sh

```
 
```
```
#!/bin/bash 
```
```
# configure the nfs server
```
```
# load parameters from prepare section
```
```
cd `dirname $0`
```
```
source prepare.sh
```
```
 
```
```
#-> start routine
```

# Edit the file /etc/exports adding the line:

```
cat >>/etc/exports << EOF 
```

/srv/nfs ${NETWORK}/${NETWORK_MASK}(rw,async,no_root_squash)

```
 
```
```
EOF
```
```
 
```
```
exportfs -a
```
```
#<- end routine
```

Proceed

To run an NFS server, the portmap service must be running. To verify that portmap is active, type the following command as root:

service portmap status

If the portmap service status is running, then
- the nfs service can be started with:service nfs start
- the nfslock service can be started with:service nfslock start.
Else execute: service portmap start and after successful report,
- the nfs service can be started with:service nfs start
- the nfslock service can be started with:service nfslock start.

By default, the nfs service does not start automatically at boot time. To configure the NFS to start up at boot time, use an initscript utility, such as /sbin/chkconfig, /sbin/ntsysv, or the Services Configuration Tool program.

administrator's script: proceed.sh

```
 
```
```
#!/bin/bash
```
```
# proceed nfs
```
```
#-> start routine
```
```
# Start the services:
```
```
/etc/init.d/portmap start 
```
```
/etc/init.d/nfs start 
```
```
/etc/init.d/nfslock start 
```

# and add them to the start configuration with:

```
chkconfig --level 35 portmap on 
```
```
chkconfig --level 35 nfs on 
```
```
chkconfig --level 35 nfslock on 
```

# If the services are not present in the level management add them with "--add" chkconfig option.

```
 
```
```
# Stop the services:
```
```
/etc/init.d/portmap stop
```
```
/etc/init.d/nfs stop
```
```
/etc/init.d/nfslock stop
```
```
 
```
```
# alternative:
```
```
/sbin/service nfs start
```
```
/sbin/service nfs stop
```
```
/sbin/service nfs restart
```
```
/sbin/service nfs condrestart
```
```
/sbin/service nfs reload
```
```
#<-end routine
```

Initial test

Verify that all necessary daemons are running. Use rpcinfo -p

# should be something like this: 
program vers proto   port
    100000    2   tcp    111  portmapper
    100000    2   udp    111  portmapper
    100024    1   udp    719  status
    100024    1   tcp    722  status
    100011    1   udp    847  rquotad
    100011    2   udp    847  rquotad
    100011    1   tcp    850  rquotad
    100011    2   tcp    850  rquotad
    100003    2   udp   2049  nfs
    100003    3   udp   2049  nfs
    100003    4   udp   2049  nfs
    100003    2   tcp   2049  nfs
    100003    3   tcp   2049  nfs
    100003    4   tcp   2049  nfs
    100021    1   udp  32775  nlockmgr
    100021    3   udp  32775  nlockmgr
    100021    4   udp  32775  nlockmgr
    100021    1   tcp  32789  nlockmgr
    100021    3   tcp  32789  nlockmgr
    100021    4   tcp  32789  nlockmgr
    100005    1   udp    863  mountd
    100005    1   tcp    866  mountd
    100005    2   udp    863  mountd
    100005    2   tcp    866  mountd
    100005    3   udp    863  mountd
    100005    3   tcp    866  mountd

administrator's script: test.sh

```
 
```
```
#!/bin/bash
```
```
# initial tests for nfs server
```
```
 
```
```
#-> start routine
```
```
rpcinfo -p
```
```
 
```

# Verify that the following packets are installed with the command:

```
echo `rpm -qa | grep nfs-utils` 
```
```
echo `rpm -qa | grep portmap`
```
```
echo `rpm -qa | grep quota`
```
```
#<- end routine
```

Update

There are some software managers (e.g. yum, yast) for operating systems which do the job well for update and new installation for packages. Please see the appropriate attributes for current task in the manuals for them (use man yum).

administrator's script: update.sh

```
 
```
```
#!/bin/bash
```
```
# update
```
```
#-> start routine
```
```
su -c 'yum update nfs-utils'
```
```
#<- end routine
```

Please open a NGI-DE ticket if you experience any Installation or Configuration problem.

Nfs/3/client

Prepare

NFS shares are mounted on the client side using the mount command. The format of the command is as follows:

mount -t <nfs-type> -o <options> <nfs_server_host>:</remote/export> </local/directory>

where <nfs-type> one from {nfs, nfs4}, <options> - a comma separated list of options for the NFS file system (refer to Common NFS Mount Options for details). Reference installation use the singe mount point, hence </remote/export> and </local/directory> are /srv/nfs.

Firewall configuration

configure TCP port 2049 (how to open port in firewall).

administrator's script: prepare.sh

```
 
```
```
#!/bin/bash
```
```
# prepare
```
```
# execute as root
```
```
# echo `su -`
```
```
 
```

# Declare the variables section ------------

# Please insert your actual configuration

# NFS_PORT="Communication port for nfs service"

```
# NFS_SERVER="adress for nfs server"
```

# NFS_SERVER_DIR="export directory on nfs server"

# NFS_MOUNT_DIR="mount directory on nfs client"

# from here ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

```
NFS_PORT="2049"
```
```
NFS_SERVER="dgiref-ifs1.fzk.de"
```
```
NFS_SERVER_DIR="/srv/nfs"
```
```
NFS_MOUNT_DIR="/srv/nfs"
```

# till here ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Install

The Linux operating system offers two methods for mounting remote file systems:

manually, with mount service (using a /etc/fstab configuration file)
automatically at boot time with the autofs service (using a main configuration file /etc/auto.master and set of auto.* files)

The reference installation use the first, manual method, hence the mount service should be installed. Probably, the operating system has nfs service installed. To be sure, the following packages can be additionally installed:

nfs-utils
portmap
quota

administrator's script: install.sh

```
 
```
```
#!/bin/bash
```
```
# install nfs client
```
```
# execute as root
```
```
# echo `su -`
```
```
#-> start routine
```
```
yum -y install mount
```
```
#<- end routine
```

Configure

With the daemons portmap, lockd, statd, mountd running NFS client should be able to mount the remote directory /srv/nfs and after creates the soft links to the mounted /srv/nfs subdirectories.

administrator's script: configure.sh

```
 
```
```
#!/bin/sh
```
```
# configure the nfs client
```
```
# load parameters from prepare section
```
```
cd `dirname $0`
```
```
source prepare.sh
```
```
 
```
```
#-> start routine
```
```
# execute as root: 
```
```
# su -
```
```
# check if the mount directory exists:
```

[ -d "${NFS_MOUNT_DIR}" ] && echo "Directory ${NFS_MOUNT_DIR} exists" || echo `mkdir -p ${NFS_MOUNT_DIR}`

```
 
```

# If the mount command works add the mounting at the start-up editing the file /etc/fstab:

echo "${NFS_SERVER}:${NFS_SERVER_DIR} ${NFS_MOUNT_DIR}  nfs nfsvers=3,bg,hard,intr,nosuid,rsize=32768,wsize=32768,rw  0 0" >> /etc/fstab

```
mount -a
```
```
#<- end routine
```

Proceed

To communicate with an NFS server, the portmap service must be running. To verify that portmap is active, type the following command as root:

service portmap status

If the portmap service status is running, then
- the nfs service can be started with:service nfs start
- the nfslock service can be started with:service nfslock start
- the mount service can be started with:service nfslock start.
Else execute: service portmap start and after successful report,
- the nfs service can be started with:service nfs start
- the nfslock service can be started with:service nfslock start
- the mount service can be started with:service nfslock start.

By default, the nfs service does not start automatically at boot time. To configure the NFS to start up at boot time, use an initscript utility, such as /sbin/chkconfig, /sbin/ntsysv, or the Services Configuration Tool program.

administrator's script: proceed.sh

```
 
```
```
#!/bin/bash
```
```
# proceed
```
```
#-> start routine
```
```
# Start the services:
```
```
/etc/init.d/portmap start 
```
```
/etc/init.d/nfs start 
```
```
/etc/init.d/nfslock start 
```

# and add them to the start configuration with:

```
chkconfig --level 35 portmap on 
```
```
chkconfig --level 35 nfs on 
```
```
chkconfig --level 35 nfslock on 
```

# If the services are not present in the level management add them with "--add" chkconfig option.

```
#<-end routine
```

Initial test

Lists the contents of the mounted directory with ls command.

WARNING: 
Possible troubles with mount
Problem 1
mount: sn05:/var/spool/pbs/server_logs failed, reason given by server: Permission denied
Possible solution: Check iptables and /etc/hosts.allow; /etc/hosts.deny
Problem 2
mount clntudp_create: RPC: Program not registered
Possible solution: chmod 644 /etc/hosts.allow

administrator's script: test.sh

```
 
```
```
#!/bin/bash
```
```
# test
```
```
# load parameters from prepare section
```
```
source prepare.sh
```
```
 
```
```
#-> start routine
```
```
ls ${NFS_MOUNT_DIR}
```
```
#check the port state
```

echo `netstat -anp | grep "${NFS_PORT}"`

```
#<- end routine
```

Update

Use the operating system package managers to implement updates, upgrades and to delete a package. Specifically for the nfs updates and patches, refer into the following online resource: http://www.linux-nfs.org/

administrator's script: update.sh

```
 
```
```
#!/bin/bash
```
```
# update
```
```
#-> start routine
```
```
su -c 'yum update nfs-utils'
```
```
#<- end routine
```

cluster:Nfs/3

Contents

Nfs/3/server

Prepare

Install

Configure

Proceed

Initial test

Update

Nfs/3/client

Prepare

Install

Configure

Proceed

Initial test

Update

Views

Personal tools

general

guidelines

reference installation

Search

internal

Toolbox