Table of Contents
Slackware Experience
Personal notes on using Slackware. Some old notes have been archived.
Do note that Slackware has a great documentation site.
Getting Slackware
The official way to do this is, of course, to get it from slackware.com.
Personally, I have getslack, a bash script based on (more accurately, a trimmed-down version of) the excellent (he termed it infamous) mirror-slackware-current.sh by Alien Bob. When going down this path, the next step would be to prepare the installation media.
Slackware Installer ISO Image
I no longer need an ISO image (refer to USB installer below). But, I have my slack2iso script (also based on Alien Bob's script) that can help in creating one using the tree downloaded by getslack.
Slackware USB Installer
Alien Bob has provided a script to make/setup/configure a USB-based Slackware installation media. I wanted to do something simpler using the existing files in the Slackware tree that I mirrored using getslack (mentioned above). So, here is how I got that working.
- Create a FAT32 partition
- use
fdiskand make sure it is bootable (bootable flag enabled) - use
mkdosfs(e.g.mkdosfs -F 32 /dev/sdb1)
- Use syslinux to provide bootloader
- create a
/linux/boot/syslinuxfolder on the USB - type
syslinux -d /linux/boot/syslinux /dev/sdb1
Note: On newer syslinux, use
-ito indicate new installation - a file
ldlinux.sysshould appear in/linux/boot/syslinux
- Copy boot facilities from Slackware tree to the media
- copy a kernel from slackware tree to
/linux/boot(I usedhuge.s) - copy
initrd.imgandmessage.txtto/linux/boot - copy
isolinux.cfgto/linux/boot/syslinuxassyslinux.cfg - edit
syslinux.cfgaccordingly (initrd, kernel params, etc.)
- Copy slackware<64> in the Slackware tree (I used a shorter folder name like slack on the USB)
And… we're done! Now we have a simple Slackware USB Installer and install it on every computer we can get our hands on!
Note: GPT Disks and EFI
Things moving to (U)EFI and GPT… slowly leaving legacy BIOS and MBR.
Instead of MBR, we use GPT partitioning scheme:
- supports bigger disk
- supports EFI booting (easier to maintain actually :p)
Partition codes are 2-bytes instead (only 1-byte on MBR's partition table). Among the common ones:
- EF00 (EFI System Partition): this is what EFI boot look for
- format FAT32
mkdosfs -F 32 -n MY1EFI /dev/sdxx
- 0700 (MS Basic Data): Windows Partition
- format NTFS
mkntfs -f -L MY1WIN /dev/sdxx
- 8300 (Linux filesystem): Linux Partition
- format EXT4
mkfs.ext4 -L MY1LIN /dev/sdxx
Once boot using EFI, efibootmgr tool can be used (available on Slackware 14.2)
- to create an entry labelled
Slackwarewith loader file named\efi\slackware\elilo.efilocated on first partition of first disk (/dev/sda1)efibootmgr -c -d /dev/sda -p 1 -L "Slackware" -l "\efi\slackware\elilo.efi"
- to delete an entry xxxx (bootnum)
efibootmgr -b xxxx -B
- to re-order boot sequence
efibootmgr -o xxxx,yyyy,zzzz
Installing Slackware
Installation notes (i.e. packages, configs).
LastUpdated20250112
Basic Install
Using Slackware installer.
- official packages (
getslack)- checkout my
getslackconfig file - without kde (AND xfce if going DE-less)
removepkg gnuchess xaos xsnowremovepkg joe nano vim-gvim slackpkg
- setup/config
- sample elilo.conf
- elilo.conf
prompt #chooser=simple timeout=50 default=Slack image=vmlinuz-huge label=SlackHuge read-only append="root=/dev/sda2 resume=/dev/sda4 vga=normal" image=vmlinuz label=Slack initrd=initrd.gz read-only append="root=/dev/sda2 resume=/dev/sda4 vga=normal"
- make sure vim does not create backups (edit
/usr/share/vim/vimrc)- or, run
vimstart(from my1shell repo)
- dmesg no longer allowed for user
- append
rc.local←echo 0 > /proc/sys/kernel/dmesg_restrict - or, run
setup_slack(from my1shell repo)
- additional packages (
getslackpack)- (alien) openjdk libreoffice libreoffice-dict-en“
- additional packages (
getslackbuild)slackware-xdm-themegeany unrarnss-mdns avahi libdaemon- actually, scripts from slackbuilds.org (commonly used)
DE-less config
This is what I do for a lean (not necessarily minimal, but trimmed to my liking) installation.
- setup acpi from my personal script
- additional packages (
getslackbuild)dmenu slock st wmnamerox-filer pmount
- custom dwm build
- using my own build script (which has personalized patches)
Updating
To maintain:
- run slack-update
- this actually runs 3 scripts (
getslack,getslackpack,getslackbuild)
- run slackpatch (if required)
- run getslackbuild build -x -i (if required)
Sample configuration files for the above scripts are here.
Using Slackware-current
This is actually NOT recommended for beginners. But, sometimes, the need to use the latest software is unavoidable and this COULD be a solution. Plus, this will add a LOT of COOL-points
Note: I have removed a section on DE-less installation since my current Slackware installations ARE, in fact, DE-less.
Note: I have also removed a section on hijacking other Linux system - this, here, turned out to be VERY similar to what needed to be done.
Installing
[LastUpdated20210620]
I need to use GTK3 version that is newer that the one on 14.2, so I tried the development version (slackware64-current). I have done the same once (pre-11), so I am aware that there can be some issues when doing this. I am happy to say that I AM writing this on a slackware64-current (15.0 beta?) installation on my laptop.
So, this is a little note to my future self (or anybody that may be find this useful DISCLAIMER: Use this at your own risk!). I am doing this while still using Devuan and I want to keep that for backup, in case things go wrong. (On a side note, the reason I use Devuan was because of the GTK3 version.) So, I have an extra partition that I have reformatted and prepared to download the stuffs I need.
- download official packages (
getslack)- create download path:
<mount-path>/home/share/slackware - create custom getslack config
.getslack - set
VERS=current - exclude kde & xfce
- setup EFI boot
bzImagein kernels/huge.s (rename tovmlinuz)initrd.imgin isolinux/ (this has the slackware setup)
- boot and run installation as usual
- DO NOT format partition (packages are there!)
- pick packages from mounted path
- manually set kernel to boot (i use huge - generic needs initramfs)
- boot newly installed slackware
- remove
gnuchessandxaospackages - make sure vim does not create backups (edit
usr/share/vim/vimrc) - allow
dmesgfor user- append
etc/rc.d/rc.local←echo 0 > /proc/sys/kernel/dmesg_restrict
- just for personal reference, some useful info on using nmcli
nmcli r[adio] wifi nmcli r[adio] wifi on nmcli d[evice] wifi list nmcli d[evice] wifi connect <ssid> password <pass> ifname <wlan0> nmcli c[onnection] show nmcli c[onnection] down <ssid> nmcli c[onnection] up <ssid>
- customize
etc/xdg/user-dirs.defaults(standard default paths) - create user
- get additional packages (
getslackpack)- luckily, alienBob's repo 'supports' current
- create custom getslackpack config
.getslackpack - (alien) openjdk libreoffice libreoffice-dict-en“
- get additional packages (
getslackbuild) - i want to use dwm
- using my own custom build script (which has personalized patches)
- my dwm xinitrc will run loginctl hibernate when battery<30% (→ what i need on my current laptop)
Updating
To maintain:
note: my libmy1slack library will detect current when etc/slackware-version has '+' suffix. this sign will disappear when -current is near to a stable release.
- run slack-update as usual
- when
-currentgoing stable, useSLACKVERS=current slack-update
- run slack-current instead of
slackpatch- when
-currentgoing stable, use-fswitch - to see removed packages, use slackview (i.e.
SLACKVERS=current slackview find –alien)
- update those installed using getslackbuild if needed
Special Needs
Sometimes these can be useful.
Slackware in chroot
20110621 I want to have a 32-bit system running in chroot environment on my Slackware64. I've used such system on Debian using schroot…
20110906 I managed to do this as published here…
20120518 Minor change to the fstab entry for dev, which needs an rbind option so that the pty inside can be valid! Discussed here.
20120524 This is now part of my slackstuff collection (now known as my1shell)… in form of a script called slackroot.
20121031 The path to the chroot installation MUST ALL BE owned by root - or else, users will get a Write failed: Broken pipe error.
TODO A how-to on creating 32-bit chroot on 64-bit Slackware using slackroot script.
- slack_chroot32.txt
- on my pure slack64 (maintained using getslack/getslackpack) $ ARCH=i686 getslack - create root filesystem using 32-bit packages # slackroot /opt/chroot32 --arch x86 --desk -x - copy user/group info from 64-bit system to chroot32 = will maintain its own login info! # preproot --init /opt/chroot32 - mount bind 'system' paths # preproot /opt/chroot32 - ssh into system to use 32-bit chroot # ssh user@127.0.0.1 - unmount bind 'system' paths # preproot --done /opt/chroot32
Slackware Multilib
Main references are here and here.
- Basically, need to have C library and compiler capable of multilib. I use
getslackpackto download required packages from Eric's multilib site. Install as instructed. - I already have a 32-bit Slackware tree downloaded using
getslackwhich I use for my 32-bit chroot installation. I usemassconvert32.shscript on this tree. Themassconvert32.shscript can be used to update as well (built packages are not rebuilt). Install as instructed. - My
slackpatchscript has been updated to handle 'blacklisted' 64-bit versions and 'upgraded' compat32 packages
Update20180903 Update20250326
Read here. I now have a more specific script to get multilib stuff (previously part of getslackpack script),
- use getslack-multilib to download alien_bob's multilib stuff
- compat32-tools glibc(&friends) gcc(&friends) [multilib packages]
- compat32 library packages [32-bit packages]
- upgrade pure-64 glibc/gcc packages counterparts
- upgradepkg –reinstall –install-new *.t?z
- note: this includes compat32-tools package (helper scripts, noarch)
- install 32-bit layer support libraries
- upgradepkg –install-new slackware64-compat32/*-compat32/*.t?z
- obviously, can be used to upgrade as well
Configure slackpatch to check/ignore multilib stuff
- dot-slackpatch
PKG_IGNORED="ffmpeg" # ignoring these standard packages => multilib! PKG_IGNORED="$PKG_IGNORED aaa_glibc-solibs" PKG_IGNORED="$PKG_IGNORED gcc gcc-brig gcc-g++ gcc-gdc gcc-gfortran" PKG_IGNORED="$PKG_IGNORED gcc-gnat gcc-go gcc-objc" PKG_IGNORED="$PKG_IGNORED glibc glibc-i18n glibc-profile"
Slackware Upgrade
For reference.
Upgrading 14.1 to 14.2
A bash shell script I used to upgrade from 14.1 to 14.2.
- upgrade_14-1_to_14-2.sh
#!/bin/bash # - upgrade 14.1 to 14.2 SLACKVERS="14.2" # setup path to slackware tree SLACKROOT=${SLACKROOT:="$(pwd)"} [ -z "$SLACKARCH" ] && [ -n "$ARCH" ] && SLACKARCH=$ARCH SLACKARCH=${SLACKARCH:="$(uname -m)"} SLACKSUFX=${SLACKSUFX:=""} [ "$SLACKARCH" == "x86_64" ] && SLACKSUFX="64" SLACKFULL=${SLACKFULL:="slackware${SLACKSUFX}"} [ -z "$SLACKVERS" ] && [ -n "$RELEASE" ] && SLACKVERS=$RELEASE SLACKVERS=${SLACKVERS:="current"} SLACKRELS=${SLACKFULL}-${SLACKVERS} SLACKPATH=${SLACKROOT}/${SLACKRELS} # step 1 upgradepkg ${SLACKPATH}/${SLACKFULL}/a/glibc-solibs-*.txz # step 2 upgradepkg ${SLACKPATH}/${SLACKFULL}/a/pkgtools-*.txz upgradepkg ${SLACKPATH}/${SLACKFULL}/a/tar-*.txz upgradepkg ${SLACKPATH}/${SLACKFULL}/a/xz-*.txz upgradepkg ${SLACKPATH}/${SLACKFULL}/a/findutils-*.txz # step 3 for dir in a ap d k l n t tcl x xap xfce ; do do_path=${SLACKPATH}/${SLACKFULL}/$dir [ ! -d $do_path ] && continue ( cd $do_path ; upgradepkg --install-new *.t?z ) done # step 4 removepkg ConsoleKit apmd bluez-hcidump cxxlibs foomatic-filters \ gnome-icon-theme imlib kdeadmin kdenetwork kdesdk kdetoys kwallet \ lesstif libelf libjpeg libxfcegui4 networkmanagement obex-data-server \ obexfs open-cobol oxygen-gtk3 phonon-mplayer phonon-xine pil portmap \ procps qca-cyrus-sasl qca-gnupg qca-ossl udev xchat xf86-input-aiptek \ xf86-video-modesetting xfce4-mixer xfce4-volumed xfwm4-themes # step 5 # - run chknew # step 7 # - check boot config
Slackware System
Useful system level notes.
Listing Packages
Simply list files in /var/log/packages (path is actually a link to /var/lib/pkgtools/packages/).
$ ls -l /var/log/packages/
List full package name (with version) only
$ for that in $(ls /var/log/packages/) ; do echo $that ; done | sort
List installed package name only
$ list=$(ls /var/log/packages/) ; for that in $list ; do that=${that%-*} ; that=${that%-*} ; that=${that%-*} ; echo $that ; done
Find installed packages that are not in Slackware tree
$ slackfull=slackware64-15.0 ; slacktree=/home/share/slackware/$slackfull ; for pkg in $(ls /var/log/packages/) ; do pkgf=$(find $slacktree/ -name "${pkg}.txz") ; [ -f "$pkgf" ] && continue ; echo "** Package '$pkg' is alien!" ; done
This can also be done using slackview script (in my1shell repo)
$ slackview find --alien
View information on specific package
$ slackview find [pkg_name]
To list all official packages without DE
$ for pack in a d k l n ap t tcl x xap ; do slackview file --name pkgs.txt --insert --pack $pack ; done
To list currently installed packages (e.g. to be used in my1live)
$ slackview file --name pkgs.txt --insert --installed
To sort packages in pkgs.txt based on software sets
$ slackview file --name pkgs.txt --sort
To remove packages in dups.txt that is/are found in pkgs.txt
$ slackview file --name pkgs.txt --dups dups.txt
Building Custom Kernel
- run shell script (getlinux)
- select version, download source
- extract at
/usr/src/(e.g. linux-4.4.199) - copy a config from
/boot# cp /boot/config-generic? > .config
- use that config
# make oldconfig
- configure build
# make menuconfig
- build the kernel
# make -j4 bzImage
- build/install modules
# make -j4 modules && make modules_install
modules_installrequires root, obviously!
- copy (as root) kernel
# cp arch/x86/boot/bzImage /boot/vmlinuz-generic-4.14.12 # cp System.map /boot/System.map-generic-4.14.12 # cp .config /boot/config-generic-4.14.12
- generate initrd if using generic
# mkinitrd -c -k 4.14.12 -f ext4 -r /dev/sda3 -m ext4 -u -o /boot/initrd.gz
- a useful initrd generator script IS available
- run
/usr/share/mkinitrd/mkinitrd_command_generator.sh - then run the generated/recommended command
- checkout the
-Poption if required
Slackware Desktop
Little tips for better Slackware desktop experience. Most of these can be used in other Linux distributions as well.
Mplayer (gmplayer) runs without video window
Double-clicking in file manager (Thunar) shows no video (audio is ok, so program is running). Running mplayer from command line is fine. Turns out gui version is gmplayer - running from console shows this error message
Failed to open VDPAU backend libvdpau_va_gl.so
So, just edit ~/.mplayer/gui.conf, find the line for vo_driver and set it to gl. Show be ok after that… at least in my case it is.
Mplayer and Xscreensaver
Note: NOT tested… just found this
Although the mplayer setting to disable xscreensaver has been selected, xscreensaver still runs!
Possible Solution: Edit ~/.mplayer/config
heartbeat-cmd="xscreensaver-command -deactivate >/dev/null 2>&1"
Login issue
When using xdm, login fails sometimes with error message … Unable to establish ICE listener …
- append
rc.local←rm -rf /tmp/.ICE-unix/*
Audio issue when using ALSA
The solution is here.
- in short, ALSA using the HDMI output as default
- use alsamixer, [F6] select the card (the one that is NOT HDMI) and make sure output is not muted
- use aplay -l to identify the card/slot# and device#
- use
aplay -D plughw:<card/slot#>,<device#> <WAV file>to test - modify
asound.confin etc or~/.asoundrcto fix this (sample below)- asound.conf
pcm.!default { type hw card <card/slot#> } ctl.!default { type hw card <card/slot#> }
Low volume when using ALSA
As seen here…
Check asound.conf in etc:
pcm.!default {
type plug
slave.pcm "softvol"
}
pcm.softvol {
type softvol
slave {
pcm "dmix"
}
control {
name "Pre-Amp"
card 0
}
min_dB -5.0
max_dB 20.0
resolution 6
}
Sometimes need type in pcm.!default to be hw
Slackware Applications
Running applications (@software) on Slackware… most are applicable to all distributions.
'Hidden' Application
There are a couple of (a few?) applications that I thought have great features but relatively unknown (I only knew about them after looking for specific solution). Here they are:
- xfig - nice app to create figures for use with latex
- xpaint - can do screen capture (
xpaint -snapshot)
Apache (web server) Setup
Modify httpd config file:
- change document root to a folder my main user have full access
- also create a link to it from my user account
- allow php to be indexed ⇒ add to directory index in dir_module
- enable php (towards the end of the file) ⇒ include mod_php.conf
- enable mod_rewrite (API server?) ⇒ LoadModule … mod_rewrite.so
- optional: set serveradmin email and servername
- optional: to only serve locally, set listen localhost:80 ???
More modifications for https:
- enable loadmodule mod_ssl.so
- enable loadmodule mod_socache_shmcb.so
- generate private key:
openssl genrsa -out privkey.pem 2048
- rsa private key with 2048-bit long modulus written to file
- generate cert:
openssl req -new -x509 -key privkey.pem -out cacert.pem -days 1095
- include httpd-ssl.conf
- modify httpd-ssl.conf accordingly…
mariadb/mysql Setup
- run
mysql_install_db
- make sure permision given to user mysql
chown -R mysql:mysql /var/lib/mysql
- start daemon
rc.mysqld - run
/usr/bin/mysql_secure_installation
- create specific database for specific app
create database app_db;
- create specific user for specific app and grant all access
grant all privileges on app_db.* to 'user_app'@'localhost' identified by 'pass_app';
- just formality, run
flush privileges;
- to additionally create specific user for specific app
create user 'user_app'@'localhost' identified by 'pass_app';
- recover root password:
- stop daemon
rc.mysqld - run
# mysqld_safe --skip-grant-tables & # mysql -u root $ mysql -uroot -p mysql> use mysql; mysql> update user set password=PASSWORD('<newpass>') where User='root'; mysql> flush privileges; mysql> exit
Backup a DB:
mysqldump -p -u user userdb > userapp-$(date +%Y%m%d%H%M%S).sql
TeXLive Install/Setup
The default tetex is usable, but it is no longer maintained and some new packages are not available.
The recommended TeX distribution is TeXLive.
- Download from here
- get Unix Installer
- extract path:
/home/share/tool/texlive
- Execute
./install-tl -gui
- install path:
/home/share/tool/texlive/YYYY(currently, YYYY=2020) - select basic, then customize (e.g. lang:en, graphics, math, etc.)
- setup environment variable
TL_VERS="2020" TL_PATH="/home/share/tool/texlive/${TL_VERS}" export PATH=${TL_PATH}/bin/x86_64-linux:$PATH export MANPATH=${TL_PATH}/texmf-dist/doc/man:$MANPATH export INFOPATH=${TL_PATH}/texmf-dist/doc/info:$INFOPATH
- setup tlmgr (tlmgr option repository ctan)
- Run manager
tlmgr --gui
- needs perl-tk
- to get collection of a package
tlmgr show <package>
rsync server
- create
rsyncd.confin etc- rsync.conf
max connections = 2 log file = /var/log/rsync.log timeout = 300 [share] comment = Shared Stuff path = /home/share read only = yes list = yes hosts allow = 192.168.3.0/24 uid = nobody gid = nobody #auth users = pub #secrets file = etc/rsyncd.secrets
- modify subnet mask address for host allow accordingly
- in
inetd.conf(etc), insertrsync stream tcp nowait root /usr/bin/rsync rsync --daemon
- in
services(etc), insertrsync 873/tcp
- create
rsyncd.secretsin etc- rsyncd.secrets
pub:pub
- start rsync daemon
/usr/bin/rsync --daemon --config=etc/rsyncd.conf
Steam on Slack64 (in chroot32)
on an x86_64 machine,
- install libtxc_dxtn package from slackbuilds.org (64-bit)
- use 'slackroot' to create 32-bit chroot environment (chroot32)
- target for desktop
- ssh into localhost to enter chroot32 as normal user
- remember to run 'preproot' (as root) prior to that
- … and 'preproot –release' when done
- install packages
- install alien_bob's steamclient package
- install libtxc_dxtn package from slackbuilds.org (32-bit)
- run 'linux32' to create an official 32-bit environment
- run 'steam -tcp'
Issues
Issues when running Slackware…
Libreoffice
- alien_bob's libreoffice cannot run
- got
unspecified application error
- found a fix in lq forum
- look for graphic driver used
$ lspci -vv | sed -n '/VGA compatible/,/^$/ p'
- e.g. for i915, do
export MESA_LOADER_DRIVER_OVERRIDE=i915
Extra Notes
Some things to note…
20210620 Slackware's CPU frequency scaling works (checkout rc.cpufreq) - just to remind myself, no need to look into this!
Admin-stuff for non-root user
To allow non-root users basic admin (poweroff,reboot,etc.), add them to power group and insert the following to sudoers.
%power ALL=(ALL) NOPASSWD:/sbin/poweroff %power ALL=(ALL) NOPASSWD:/sbin/reboot %power ALL=(ALL) NOPASSWD:/usr/sbin/pm-suspend %power ALL=(ALL) NOPASSWD:/usr/sbin/pm-hibernate %power ALL=(ALL) NOPASSWD:/usr/sbin/pm-powersave
Note: The pm-* binaries (pm-utils) are no longer available on Slackware 15.0
Note20250112: use loginctl on Slackware 15.0 - no need to add sudoers
Multicast DNS @ Zeroconf
- install nss-mdns (requires avahi, which requires libdaemon)
- save nss config (in etc)
mv nsswitch.conf nsswitch.conf-orig
- use provided nss config
cp nsswitch.conf-mdns nsswitch.conf
- run daemon at startup
rc.avahidaemonandrc.avahidnsconfd
Note20260720: latest install seems to NOT need to switch@edit nssswitch.conf
NIS on Slackware
Got this from here.
Slackware-NIS-mini-HOWTO
How to set up a NIS server and configure NIS clients on Slackware Linux
14-Dec-2000
David Cantrell <david@slackware.com
==============
THE NIS SERVER
==============
STEP 1: Decide which machine will be the NIS server
----------------------------------------------------
This is usually not a very involving process, but does require some
thought. You'll want to use a machine that is on most of the time and
does not need to reboot into other operating systems. If you have an
existing fileserver, that would be a good choice for your NIS server.
STEP 2: Decide on what you want NIS to serve
---------------------------------------------
NIS is fairly complex. It allows you to distribute all kinds of
information. Things like usernames, passwords, groups, hostnames, and
network services that should be running. The most common use is to use
NIS as a common user logon authetication system. The second most common
use is to have it share NFS automount maps so that users logging in on a
system will have their home directory automounted for them.
In this mini-HOWTO, I will explain how to set up your NIS server and
clients for passwd/group sharing and home directory sharing.
STEP 3: Move home directories to sharing location
--------------------------------------------------
In order to make the home directory NFS shares work seamlessly across the
NIS server and clients, I move them to /export/home. This is the location
they will be mounted on the client, so I make them exist there on the
server as well. You might not have /export on your server, so let's make
it:
mkdir -p /export
Now make the home directory location:
mkdir /export/home
Now move all your home directories over to /export/home. You should leave
"ftp" in /home, as that's where Slackware installs it and it is sometimes
site specific. These commands should move all home directories to
/export/home and leave ftp in /home:
mv /home/* /export/home
mv /export/home/ftp /home
STEP 4: Edit home directory paths on the server
------------------------------------------------
Since home directories are now in /export/home, you need to edit
/etc/passwd on the server to reflect those locations. If you have a user
called "david", you will want to change his home directory to
/export/home/david. Do this for all users on the system.
STEP 5: Write the automount maps
---------------------------------
I usually make two automount maps for the home directories. One is called
auto.master and controls how the automounter works. The other is called
auto.export and will describe the /export map.
Edit /etc/auto.master and add this to it:
#
# auto.master - Master map file for NFS automounter
#
/export auto.export --timeout 60
Now edit /etc/auto.export and add this to it:
#
# auto.export - NFS automounter configuration file
#
# Home directories
home -fstype=nfs <NFS/NIS server>:/export/home
Of course, replace <NFS/NIS server> with the hostname of your NFS/NIS
server.
STEP 6: Edit the NFS exports file on the server
------------------------------------------------
Now we need to make sure that /export is shared from the server. Edit
/etc/exports and make sure this line is present:
/export (rw,no_root_squash)
Save the file and restart the NFS server daemons:
killall -HUP rpc.nfsd rpc.mountd
STEP 7: Pick a name for your NIS domain
----------------------------------------
Each NIS domain needs to have a unique name. Do not use the same name as
your DNS domain. This should be something different. I will use "mynis"
in this example.
Edit /etc/defaultdomain and add the name of your NIS domain:
mynis
Save and exit. The name of your NIS domain shoulw be the only thing in
that file. Once you've done that, set the NIS domain with this command:
nisdomainname `cat /etc/defaultdomain`
STEP 8: Edit /var/yp/Makefile
------------------------------
The NIS server has a Makefile that controls what maps are served out by
NIS. Each time you make a change (such as adding a user), you should go
into /var/yp and type "make" to update the NIS information.
In this file we need to edit the all: target and specify the stuff we want
to share. Here's what your all: target should look like:
all: passwd group hosts rpc services netid protocols netgrp mail \
shadow auto.master auto.export
Now find the line that looks like this:
AUTO_LOCAL = $(YPSRCDIR)/auto.local
We want to add this line just after that line:
AUTO_EXPORT = $(YPSRCDIR)/auto.export
Now find the auto.local: target. We need to add this target just after
that target:
auto.export: $(AUTO_EXPORT) $(YPDIR)/Makefile
@echo "Updating $@..."
-@sed -e "/^#/d" -e s/#.*$$// $(AUTO_EXPORT) | $(DBLOAD) \
-i $(AUTO_EXPORT) -o $(YPMAPDIR)/$@ - $@
-@$(NOPUSH) || $(YPPUSH) -d $(DOMAIN) $@
Save the Makefile and exit.
STEP 9: Edit /var/yp/securenets
--------------------------------
Edit /var/yp/securenets and change the "0.0.0.0 0.0.0.0" line to
something like this:
255.255.255.0 192.168.1.0
You should specify your network address and submask there, this is only
an example. This example restricts NIS access to any machine on the
192.168.1 class C subnet.
STEP 10: Create an /etc/ypserv.conf file
-----------------------------------------
The only line you need in this file is:
dns: yes
This line tells ypserv to resolve hostnames using a DNS server instead of
relying on NIS hostname maps (which you probably aren't using). Save and
exit.
STEP 11: Start ypbind
----------------------
Start ypbind:
/usr/sbin/ypbind
Use ps to make sure it started and is running properly. You will probably
see a master and slave process for it.
STEP 12: Start ypserv
----------------------
Start ypserv:
/usr/sbin/ypserv
Use ps to make sure it started and is running properly.
STEP 13: Initialize the NIS maps
---------------------------------
We need to initialize the NIS maps with this command:
/usr/lib/yp/ypinit -m
Fill in the blanks for all machines on your network that will be acting as
NIS servers. At this point, you probably want to fire up rpc.yppasswdd to
make sure that users on NIS clients can run chfn and passwd.
STEP 14: Done, making sure NIS stuff loads at bootup
-----------------------------------------------------
Your NIS server is now up and running. You should edit /etc/rc.d/rc.inet2
and make sure the following block is uncommented:
# Setting up NIS:
# (NOTE: For detailed information about setting up NIS, see the documentation
# in /usr/doc/yp-tools, /usr/doc/ypbind, and /usr/doc/ypserv)
#
# First, we must set the NIS domainname. NOTE: this is not
# necessarily the same as your DNS domainname, set in
# /etc/resolv.conf! The NIS domainname is the name of a domain
# served by your NIS server.
if [ -r /etc/defaultdomain ]; then
nisdomainname `cat /etc/defaultdomain`
fi
# Then, we start up ypbind. It will use broadcast to find a server.
if [ -d /var/yp ] ; then
echo -n " ypbind"
${NET}/ypbind
fi
# If you are the NIS master server for the NIS domain, then
# you must run rpc.yppasswdd, which is the RPC server that
# lets users change their passwords.
if [ -x ${NET}/rpc.yppasswdd ]; then
echo -n " yppasswdd"
${NET}/rpc.yppasswdd
fi
It's probably adviseable to run rpc.yppasswdd with this line:
${NET}/rpc.yppasswdd -e chsh -e chfn
This ensures that chfn and chsh work. You also need to add this after the
ypbind if block:
if [ -x ${NET}/ypserv ]; then
echo -n " ypserv"
${NET}/ypserv
fi
The ypserv program needs to run on the NIS server.
==============
THE NIS CLIENT
==============
STEP 1: Set the NIS domain
---------------------------
Edit /etc/defaultdomain just like you did on the server and enter the name
of your NIS domain.
STEP 2: Edit /etc/yp.conf
--------------------------
Open /etc/yp.conf in a text editor and make sure this line is present:
ypserver <IP or hostname of the NIS server>
Specify either the IP address or hostname of your NIS server. If you
don't have working hostnames, you must use the IP address here.
STEP 3: Set the NIS domain
---------------------------
Set the NIS domain by typing this command:
nisdomainname `cat /etc/defaultdomain`
STEP 4: Run ypbind
-------------------
Start ypbind with this command:
/usr/sbin/ypbind
STEP 5: Test the NIS connection
--------------------------------
See if the NIS server is working with this command:
ypcat passwd.byname
If you see the /etc/passwd entries from the NIS server, you're good to go.
Otherwise, go back to the NIS server section and see if you missed any
steps.
STEP 6: Edit /etc/nsswitch.conf
--------------------------------
You need to edit /etc/nsswitch.conf and tell it to use NIS for passwd,
shadow, and group. Change those entries to:
db files nis
Change the automount entry to:
files nis
STEP 7: Enable NIS logins in /etc/passwd and /etc/group
--------------------------------------------------------
Edit /etc/passwd and add this line to the end:
+:::::
Edit /etc/group and add this line to the end:
+:::
STEP 8: Create the /export directory for home directory mounting
-----------------------------------------------------------------
Make sure you have that directory:
mkdir -p /export
STEP 9: Create an rc.autofs script
-----------------------------------
The included rc.autofs script should be moved into /etc/rc.d. Edit
rc.local and add this line:
/etc/rc.d/rc.autofs
STEP 10: Done, enable NIS stuff at bootup
------------------------------------------
Your NIS client is now configured. Edit /etc/rc.d/rc.inet2 and uncomment
this block:
# Setting up NIS:
# (NOTE: For detailed information about setting up NIS, see the documentation
# in /usr/doc/yp-tools, /usr/doc/ypbind, and /usr/doc/ypserv)
#
# First, we must set the NIS domainname. NOTE: this is not
# necessarily the same as your DNS domainname, set in
# /etc/resolv.conf! The NIS domainname is the name of a domain
# served by your NIS server.
if [ -r /etc/defaultdomain ]; then
nisdomainname `cat /etc/defaultdomain`
fi
# Then, we start up ypbind. It will use broadcast to find a server.
if [ -d /var/yp ] ; then
echo -n " ypbind"
${NET}/ypbind
fi
