“UbuntuHelp:UbuntuOnCluster”的版本间的差异

2007年11月30日 (五) 21:52的版本

文章出处:

https://help.ubuntu.com/community/UbuntuOnCluster

点击翻译:

English

请不要直接编辑翻译本页，本页将定期与来源同步。

This page documents the process of automatically installing ubuntu on a cluster of machines. This means one machine will be setup manually as install server and all others will be installed when booting. The version used here is hoary. As installer I used only the debian-installer (The default Ubuntu installer). So no tricks like FAI or kickstart are needed. Prerequisites on the server:

Ubuntu-base install
An ubuntu .iso
nat active on the server example or all machines must receive a public IP address by dhcp
Several packages need to be installed

First you will have to know the mac addresses of all machines, so they will get unique and constant IP addresses and hostnames. For me this was easy since there was a list of them available. Now install the packages dhcp3-server and tftpd-hpa Once you have this list, you can edit your /etc/dhcp3/dhcpd.conf, an example can be found here; it's self-explaining. PXE booting requires that the .iso file is mounted locally, I mounted it under /var/lib/tftpboot/ubuntu/

mkdir /var/lib/tftpboot/ubuntu
echo '/data/ubuntu-5.04-install-i386.iso /var/lib/tftpboot/ubuntu/ auto defaults,loop 0 0' >> /etc/fstab
mount -a

Next step is setting up the PXE config. I created two files: one for installing and one for booting from the local disk (ie: booting the installed system. Create /var/lib/tftpboot/pxelinux.cfg and put these files there: default localboot As you can see, the default action is to run the installer. You can also set a few options in here (the kernel command line) To save some space on the kernel command line (space is limited), create symlinks to relevant files:

cd /var/lib/tftpboot
ln -s ubuntu/install/netboot/ubuntu-installer/i386/initrd.gz
ubuntu/install/netboot/ubuntu-installer/i386/linux
ubuntu/install/netboot/pxelinux.0
ubuntu/install/netboot/ubuntu-installer/

You can see that in the example config files these symlinks are used.

Stage 2: Setting up nis and nfs

For cluster machines, nis and nfs are usually used to share login information and parts of the filesystem. So you need to install both on the server. You will need the packages nis nfs-kernel-server. Note: the nis package (not nis itself) is quite buggy, it will try to start ypbind even though you did not tell it to. It also completely ignores preseed, so in a following step we will create a new version of this package. For now it will do, you just have to wait a bit for ypbind to time out. When the nis setup asks for a domain, pick one you like. As soon as ypbind times out, stop nis again with

invoke-rc.d nis stop

Now you need to edit /etc/default/nis and enable the nis server (example). You also need to initialize the nis database with

/usr/lib/yp/ypinit -m

You can now start the nis services again

invoke-rc.d nis start

For NFS, you need to edit /etc/exports in order to export required parts of the filesystem. On my cluster I chose to export /home and /data, so the exports file looks like

# /etc/exports: the access control list for filesystems which may be exported
#               to NFS clients.  See exports(5).
/home   192.168.0.0/255.255.0.0(rw,async)
/data   192.168.0.0/255.255.0.0(rw,async)

Now restart the nfs server.

invoke-rc.d nfs restart

Stage 3: Setting up local mirror and proxy

Installing from a local mirror and using an http proxy for the rest greatly improves the speed of subsequent installs. Since apt-proxy is quite broken, I chose to use squid as a generic http proxy. For the recovering the reboot part I also needed php, so I installed that too. (A CGI script would have worked just fine here, but i'm more familiar with php) You will need the packages: apache2 libapache2-mod-php4 squid My cluster has only one external IP address, so only the master server is connected to the internet. The other machines are connected only to the master (and via NAT they can reach the net). Because I don't want to be a public proxy, I told the squid installer to only listen on eth1 (the internal interface). You can tell apache to do so to by editing /etc/apache2/ports.conf.

invoke-rc.d apache2 stop
echo 'Listen 192.168.0.1:80' > /etc/apache2/ports.conf
echo 'Listen 127.0.0.1:80' >> /etc/apache2/ports.conf
invoke-rc.d apache start

You should also edit the squid config. An example can be fond on my homepage. Now make the apache server an ubuntu archive by symlinking into the .iso

ln -s /var/lib/tftpboot/ubuntu/ubuntu /var/www/ubuntu

Stage 4: preseed

Having everything in place on the server, we can now take care of the client configuration. The tftpboot will launch the ubuntu installer. This installer usually asks questions, but the answers can be preseeded in a so called preseed file. A preseed file that answers all default questions and installs ubuntu-base and openssh-server can once again be found on my homepage here. Points for possible changes: language, package selection (which is an aptitude pattern), first-created user account and partition manager setup. Go over these settings and adapt them to your needs.

Stage 5: Surviving the reboot

The installer reboots after the basic install, which means that the installer will be launched again. Of course you don't want this, which is why I created a so called registration system. As you can see in the preseed file, the preseed/late-command has been set to wget http://192.168.0.1/register.php. This does nothing on the client side, but the php script creates a PXE boot file for this machine which instructs it to boot from the local drive. If you want to reinstall a certain machine all you have to do is remove the associated PXE boot file and it will use the default again. In order for this to work, the www-data user must have write access to /var/lib/tftpboot/pxelinux.cfg

chown :www-data /var/lib/tftpboot/pxelinux.cfg
chmod g+w $_

This is the register.php script:

<?php
function _dechex($x) { return sprintf("%02s",dechex($x)); }
$host = strtoupper(implode('',array_map(_dechex,explode('.',$_SERVER['REMOTE_ADDR']))));
copy("/var/lib/tftpboot/pxelinux.cfg/localboot", "/var/lib/tftpboot/pxelinux.cfg/$host");
?>

Stage 6: postinstall

The postinstall script can be used to do anything you like. I used it to install a correct nis package on the clients, create a correct NIS config and more bootstrapping things. Creating a correct NIS package is done as follows (for breezy this won't be neccessary: the package has been fixed).

mkdir nispackage
cd nispackage
apt-get source nis
cd nis-3.12/debian

Open the file postinst
comment out line 61 (the one with db_input
just before line 64 (the one with if [ "$RET add a line containing RET=domain where domain is the NIS domain you chose
comment out line 106 (The one with db_text)

Now enter the following sequence of commands:

cd ..
apt-get build-dep nis
dpkg-buildpackage

Copy the newly generated .deb (to be found in the nispackage folder) file to a location where the clients can retrieve them (either with wget or an nfs mount) The postinstallscript itself should be placed in /var/www Mine looks like this:

#!/bin/bash
# Step 1: Initial mount
mkdir /data
mount -t nfs 192.168.0.1:/data /data -o rw,soft,bg,rsize=32768,wsize=32768,tcp,timeo=600,intr
# Step 2: Install and configure NIS
aptitude -y install portmap libslp1
dpkg -i /data/nis_3.12-3_i386.deb
cp /data/nsswitch.conf /etc/nsswitch.conf
echo '+::::::' >> /etc/passwd
echo '+::::::::' >> /etc/shadow
echo '+:::' >> /etc/group
# Repair failing boot sequence (nis with nfs error)
DIR=`pwd`
cd /etc/rc2.d
ln -s ../init.d/mountnfs.sh S20mountnfs.sh
cd $DIR
# Step 2: Correct mounting
umount /data
echo 'enterprise:/home  /home   nfs     rw,soft,bg,rsize=32768,wsize=32768,tcp,timeo=600,intr' >> /etc/fstab
echo 'enterprise:/data  /data   nfs     rw,soft,bg,rsize=32768,wsize=32768,tcp,timeo=600,intr' >> /etc/fstab
mount -a
# Step 3: Correct sources.list and update
cp /data/sources.list /etc/apt/sources.list
aptitude update
aptitude -y upgrade
# Step 4: Installing other things and remove the installer log which contains a cleartext password
aptitude -y install linux-686-smp manpages-dev
rm /var/log/installer/debconf-seed
rm /var/log/installer/cdebconf/questions.dat
# Step 5: Enable scheduled scripts
echo '0 *    * * *  root   test -x /data/upgrade && /data/upgrade' >> /etc/crontab
# Step 6: reboot, we installed a new kernel
reboot

Downloads: example nsswitch.conf sources.list with local mirror the post-install script

@@ 第2行： / 第2行： @@
 {{Languages|UbuntuHelp:UbuntuOnCluster}}
 This page documents the process of automatically installing ubuntu on a cluster of machines. This means one machine will be setup manually as install server and all others will be installed when booting. The version used here is hoary. As installer I used only the debian-installer (The default Ubuntu installer). So no tricks like FAI or kickstart are needed.
 Prerequisites on the server:
 * Ubuntu-base install
@@ 第8行： / 第7行： @@
 * nat active on the server [http://kaarsemaker.net/files/ubuntu-cluster/enable_nat example] or all machines must receive a public IP address by dhcp
 * Several packages need to be installed
 === Stage 1: Preparing DHCP & PXE booting ===
 First you will have to know the mac addresses of all machines, so they will get unique and constant IP addresses and hostnames. For me this was easy since there was a list of them available.
 Now install the packages <code><nowiki>dhcp3-server</nowiki></code> and <code><nowiki>tftpd-hpa</nowiki></code>
 Once you have this list, you can edit your /etc/dhcp3/dhcpd.conf, an example can be found [http://kaarsemaker.net/files/ubuntu-cluster here]; it's self-explaining.
 PXE booting requires that the .iso file is mounted locally, I mounted it under /var/lib/tftpboot/ubuntu/
 <pre><nowiki>mkdir /var/lib/tftpboot/ubuntu
 echo '/data/ubuntu-5.04-install-i386.iso /var/lib/tftpboot/ubuntu/ auto defaults,loop 0 0' >> /etc/fstab
 mount -a</nowiki></pre>
 Next step is setting up the PXE config. I created two files: one for installing and one for booting from the local disk (ie: booting the installed system. Create /var/lib/tftpboot/pxelinux.cfg and put these files there: [http://kaarsemaker.net/files/ubuntu-cluster/default default]
 [http://kaarsemaker.net/files/ubuntu-cluster/localboot localboot]
 As you can see, the default action is to run the installer. You can also set a few options in here (the kernel command line) To save some space on the kernel command line (space is limited), create symlinks to relevant files:
 <pre><nowiki>cd /var/lib/tftpboot
 ln -s ubuntu/install/netboot/ubuntu-installer/i386/initrd.gz
@@ 第34行： / 第24行： @@
 ubuntu/install/netboot/ubuntu-installer/
 </nowiki></pre>
 You can see that in the example config files these symlinks are used.
 === Stage 2: Setting up nis and nfs ===
 For cluster machines, nis and nfs are usually used to share login information and parts of the filesystem. So you need to install both on the server. You will need the packages <code><nowiki>nis nfs-kernel-server</nowiki></code>.
 Note: the nis package (not nis itself) is quite buggy, it will try to start ypbind even though you did not tell it to. It also completely ignores preseed, so in a following step we will create a new version of this package. For now it will do, you just have to wait a bit for ypbind to time out.
 When the nis setup asks for a domain, pick one you like. As soon as ypbind times out, stop nis again with
 <pre><nowiki>
 invoke-rc.d nis stop</nowiki></pre>
 Now you need to edit /etc/default/nis and enable the nis server [http://kaarsemaker.net/files/ubuntu-cluster/nis (example)]. You also need to initialize the nis database with
 <pre><nowiki>
 /usr/lib/yp/ypinit -m</nowiki></pre>
 You can now start the nis services again
 <pre><nowiki>
 invoke-rc.d nis start</nowiki></pre>
 For NFS, you need to edit /etc/exports in order to export required parts of the filesystem. On my cluster I chose to export /home and /data, so the exports file looks like
 <pre><nowiki>
 # /etc/exports: the access control list for filesystems which may be exported
@@ 第65行： / 第43行： @@
 /home  192.168.0.0/255.255.0.0(rw,async)
 /data  192.168.0.0/255.255.0.0(rw,async)</nowiki></pre>
 Now restart the nfs server.
 <pre><nowiki>
 invoke-rc.d nfs restart</nowiki></pre>
 === Stage 3: Setting up local mirror and proxy ===
 Installing from a local mirror and using an http proxy for the rest greatly improves the speed of subsequent installs. Since apt-proxy is quite broken, I chose to use squid as a generic http proxy. For the recovering the reboot part I also needed php, so I installed that too. (A CGI script would have worked just fine here, but i'm more familiar with php)
 You will need the packages: <code><nowiki>apache2 libapache2-mod-php4 squid</nowiki></code>
 My cluster has only one external IP address, so only the master server is connected to the internet. The other machines are connected only to the master (and via NAT they can reach the net). Because I don't want to be a public proxy, I told the squid installer to only listen on eth1 (the internal interface). You can tell apache to do so to by editing /etc/apache2/ports.conf.
 <pre><nowiki>invoke-rc.d apache2 stop
 echo 'Listen 192.168.0.1:80' > /etc/apache2/ports.conf
 echo 'Listen 127.0.0.1:80' >> /etc/apache2/ports.conf
 invoke-rc.d apache start</nowiki></pre>
 You should also edit the squid config. An [http://kaarsemaker.net/files/ubuntu-cluster/squid.conf example] can be fond on my homepage.
 Now make the apache server an ubuntu archive by symlinking into the .iso
 <pre><nowiki>
 ln -s /var/lib/tftpboot/ubuntu/ubuntu /var/www/ubuntu</nowiki></pre>
 === Stage 4: preseed ===
 Having everything in place on the server, we can now take care of the client configuration. The tftpboot will launch the ubuntu installer. This installer usually asks questions, but the answers can be preseeded in a so called preseed file.
 A preseed file that answers all default questions and installs ubuntu-base and openssh-server can once again be found on my homepage [http://kaarsemaker.net/files/ubuntu-cluster/preseed here].
 Points for possible changes: language, package selection (which is an aptitude pattern), first-created user account and partition manager setup. Go over these settings and adapt them to your needs.
 === Stage 5: Surviving the reboot ===
 The installer reboots after the basic install, which means that the installer will be launched again. Of course you don't want this, which is why I created a so called registration system. As you can see in the preseed file, the preseed/late-command has been set to <code><nowiki>wget http://192.168.0.1/register.php</nowiki></code>. This does nothing on the client side, but the php script creates a PXE boot file for this machine which instructs it to boot from the local drive. If you want to reinstall a certain machine all you have to do is remove the associated PXE boot file and it will use the default again.
 In order for this to work, the <code><nowiki>www-data</nowiki></code> user must have write access to /var/lib/tftpboot/pxelinux.cfg
 <pre><nowiki>
 chown :www-data /var/lib/tftpboot/pxelinux.cfg
 chmod g+w $_</nowiki></pre>
 This is the register.php script:
 <pre><nowiki>
@@ 第115行： / 第75行： @@
 copy("/var/lib/tftpboot/pxelinux.cfg/localboot", "/var/lib/tftpboot/pxelinux.cfg/$host");
 ?></nowiki></pre>
 === Stage 6: postinstall ===
 The postinstall script can be used to do anything you like. I used it to install a correct nis package on the clients, create a correct NIS config and more bootstrapping things.
 Creating a correct NIS package is done as follows (for breezy this won't be neccessary: the package has been fixed).
 <pre><nowiki>
 mkdir nispackage
@@ 第131行： / 第87行： @@
 * just before line 64 (the one with <code><nowiki> if [ "$RET</nowiki></code> add a line containing RET=domain where domain is the NIS domain you chose
 * comment out line 106 (The one with <code><nowiki>db_text</nowiki></code>)
 Now enter the following sequence of commands:
 <pre><nowiki>
@@ 第137行： / 第92行： @@
 apt-get build-dep nis
 dpkg-buildpackage</nowiki></pre>
 Copy the newly generated .deb (to be found in the nispackage folder) file to a location where the clients can retrieve them (either with wget or an nfs mount)
 The postinstallscript itself should be placed in /var/www
 Mine looks like this:
 <pre><nowiki>
 #!/bin/bash
 # Step 1: Initial mount
 mkdir /data
 mount -t nfs 192.168.0.1:/data /data -o rw,soft,bg,rsize=32768,wsize=32768,tcp,timeo=600,intr
 # Step 2: Install and configure NIS
 aptitude -y install portmap libslp1
@@ 第157行： / 第107行： @@
 echo '+::::::::' >> /etc/shadow
 echo '+:::' >> /etc/group
 # Repair failing boot sequence (nis with nfs error)
 DIR=`pwd`
@@ 第163行： / 第112行： @@
 ln -s ../init.d/mountnfs.sh S20mountnfs.sh
 cd $DIR
 # Step 2: Correct mounting
 umount /data
@@ 第169行： / 第117行： @@
 echo 'enterprise:/data  /data  nfs    rw,soft,bg,rsize=32768,wsize=32768,tcp,timeo=600,intr' >> /etc/fstab
 mount -a
 # Step 3: Correct sources.list and update
 cp /data/sources.list /etc/apt/sources.list
 aptitude update
 aptitude -y upgrade
 # Step 4: Installing other things and remove the installer log which contains a cleartext password
 aptitude -y install linux-686-smp manpages-dev
 rm /var/log/installer/debconf-seed
 rm /var/log/installer/cdebconf/questions.dat
 # Step 5: Enable scheduled scripts
 echo '0 *    * * *  root  test -x /data/upgrade && /data/upgrade' >> /etc/crontab
 # Step 6: reboot, we installed a new kernel
 reboot

个人工具

“UbuntuHelp:UbuntuOnCluster”的版本间的差异 - Ubuntu中文

搜索

导航

编辑

工具