source: branches/fc13-dev/server/doc/ @ 1677

Last change on this file since 1677 was 1677, checked in by ezyang, 13 years ago
More updates.
File size: 15.1 KB
1# This document is a how-to for installing a Fedora server.
2# It is semi-vaguely in the form of a shell script, but is not really
3# runnable as it stands.
5set -e -x
7# Some commands should be run as the scripts-build user, not root.
9alias asbuild="sudo -u scripts-build"
11# Old versions of this install document advised setting
12# NSS_NONLOCAL_IGNORE=1 anytime you're setting up anything, e.g. using
13# yum, warning that useradd will query LDAP in a stupid way that makes
14# it hang forever.  As of Fedora 13, this does not seem to be a problem,
15# so it's been removed from the instructions.  If an install is hanging,
16# though, try adding NSS_NONLOCAL_IGNORE.
18# This is actually just "pick an active scripts server".  It can't be
19# because our networking config points that domain
20# at localhost, and if our server is not setup at that point things
21# will break.
24# 'branch' is the current svn branch you are on.  You want to
25# use trunk if your just installing a new server, and branches/fcXX-dev
26# if your preparing a server on a new Fedora release.
29# 'server' is the public hostname of your server, for SCP'ing files
30# to and from.
33# Start with a Scripts kickstarted install of Fedora (install-fedora)
35# Take updates, reboot if there's a kernel update.
37    yum update
39# Get rid of network manager
40    yum remove NetworkManager
42# Check out the svn repository. Configure svn not to cache
43# credentials.
45# Copy over root's dotfiles from one of the other machines.
46# Perhaps a useful change is to remove the default aliases
47    cd /root
48    ls -l .bashrc
49    ls -l .ldapvirc
50    ls -l .screenrc
51    ls -l .ssh
52    ls -l .vimrc
53    ls -l .k5login
54    # Trying to scp from server to server won't work, as scp
55    # will attempt to negotiate a server-to-server connection.
56    # Instead, scp to your trusted machine as a temporary file,
57    # and then push to the other server
58scp -r root@$source_server:~/{.bashrc,.ldapvirc,.screenrc,.ssh,.vimrc,.k5login} .
59scp -r {.bashrc,.ldapvirc,.screenrc,.ssh,.vimrc,.k5login} root@$server:~
61# Install the initial set of credentials (to get Kerberized logins once
62# krb5 is installed).  Otherwise, SCP'ing things in will be annoying.
63#   o You probably installed the machine keytab long ago
64    ls -l /etc/krb5.keytab
65#     Use ktutil to combine the host/ and
66#     host/ keys with host/ in
67#     the keytab.  Do not use 'k5srvutil change' on the combined keytab
68#     or you'll break the other servers. (real servers only).  Be
69#     careful about writing out the keytab: if you write it to an
70#     existing file the keys will just get appended.  The correct
71#     credential list should look like:
72#       ktutil:  l
73#       slot KVNO Principal
74#       ---- ---- ---------------------------------------------------------------------
75#          1    5 host/
76#          2    3 host/
77#          3    2      host/
78#   o Replace the ssh host keys with the ones common to all scripts servers (real servers only)
79    ls -l /etc/ssh/*key*
80#     You can do that with:
81scp root@$source_server:/etc/ssh/*key* .
82scp *key* root@$server:/etc/ssh/
83    service sshd reload
85# Check out the scripts /etc configuration
86    # backslash to make us not use the alias
87    cd /root
88    \cp -a etc /
89    chmod 0440 /etc/sudoers
91# NOTE: You will have just lost DNS resolution and the ability
92# to do password SSH in.  If you managed to botch this step without
93# having named setup, you can do a quick fix by frobbing /etc/resolv.conf
94# with a non address for the DNS server.  Be sure to revert it once
95# you have named.
97# NOTE: You can get password SSH back by editing /etc/ssh/sshd_config (allow
98# password auth) and /etc/pam.d/sshd (comment out the first three auth
99# lines).  However, you should have the Kerberos credentials in place
100# so as soon as you install the full set of Scripts packages, you'll get
101# Kerberized logins.
103# Make sure network is working.  If this is a new server name, you'll
104# need to add it to /etc/hosts and
105# /etc/sysconfig/network-scripts/route-eth1.  Kickstart should have
106# configured eth0 and eth1 correctly; use service network restart
107# to add the new routes in route-eth1.
108    service network restart
109    route
110    ifconfig
111    cat /etc/hosts
112    cat /etc/sysconfig/network-scripts/route-eth1
114# This is the point at which you should start updating scriptsified
115# packages for a new Fedora release.  Consult 'upgrade-tips' for more
116# information.
117    yum install -y scripts-base
118    # Some of these packages are naughty and clobber some of our files
119    cd /etc
120    svn revert resolv.conf hosts sysconfig/openafs
122# Replace rsyslog with syslog-ng by doing:
123    rpm -e --nodeps rsyslog
124    yum install -y syslog-ng
125    chkconfig syslog-ng on
127# Fix the openafs /usr/vice/etc <-> /etc/openafs mapping.
128    echo "/afs:/usr/vice/cache:10000000" > /usr/vice/etc/cacheinfo
129    echo "" > /usr/vice/etc/ThisCell
131# [TEST SERVER] If you're installing a test server, this needs to be
132# much smaller; the max filesize on XVM is 10GB.  Pick something like
133# 500000. Also, some of the AFS parameters are kind of retarded (and if
134# you're low on disk space, will actually exhaust our inodes).  Edit
135# these parameters in /etc/sysconfig/openafs
137# Test that zephyr is working
138    chkconfig zhm on
139    service zhm start
140    echo 'Test!' | zwrite -d -c scripts -i test
142# Install the full list of RPMs that users expect to be on the
143# servers.
144rpm -qa --queryformat "%{Name}.%{Arch}\n" | sort > packages.txt
145# arrange for packages.txt to be passed to the server, then run:
146# --skip-broken will (usually) prevent you from having to sit through
147# several minutes of dependency resolution until it decides that
148# it can't install /one/ package.
149    yum install -y --skip-broken $(cat packages.txt)
151# Check which packages are installed on your new server that are not
152# in the snapshot, and remove ones that aren't needed for some reason
153# on the new machine.  Otherwise, aside from bloat, you may end up
154# with undesirable things for security, like sendmail.
155    rpm -qa --queryformat "%{Name}.%{Arch}\n" | grep -v kernel | sort > newpackages.txt
156    diff -u packages.txt newpackages.txt | grep -v kernel | less
157    # here's a cute script that removes all extra packages
158    yum erase -y $(grep -Fxvf packages.txt newpackages.txt)
160# We need an upstream version of cgi which we've packaged ourselves, but
161# it doesn't work with the haskell-platform package which expects
162# explicit versions.  So temporarily rpm -e the package, and then
163# install it again after you install haskell-platform.  [Note: You
164# probably won't need this in Fedora 15 or something, when the Haskell
165# Platform gets updated.]
166    rpm -e ghc-cgi-devel ghc-cgi
167    yum install -y haskell-platform
168    yumdownloader ghc-cgi
169    yumdownloader ghc-cgi-devel
170    rpm -i ghc-cgi*1.8.1*.rpm
172# Check out the scripts /usr/vice/etc configuration
173    cd /root/vice
174    \cp -a etc /usr/vice
176# Install the full list of perl modules that users expect to be on the
177# servers.
178    cd /root
179    export PERL_MM_USE_DEFAULT=1
180    cpan # this is interactive, enter the next two lines
181        o conf prerequisites_policy follow
182        o conf commit
183# on a reference server
184perldoc -u perllocal | grep head2 | cut -f 3 -d '<' | cut -f 1 -d '|' | sort -u | perl -ne 'chomp; print "notest install $_\n" if system("rpm -q --whatprovides \"perl($_)\" >/dev/null 2>/dev/null")' > perl-packages.txt
185# arrange for perl-packages.txt to be transferred to server
186    cat perl-packages.txt | perl -MCPAN -e shell
188# Install the Python eggs and Ruby gems and PEAR/PECL doohickeys that are on
189# the other servers and do not have RPMs.
190# The general mode of operation will be to run the "list" command
191# on both servers, see what the differences are, check if those diffs
192# are packaged up as rpms, and install them (rpm if possible, native otherwise)
193# - Look at /usr/lib/python2.6/site-packages and
194#           /usr/lib64/python2.6/site-packages for Python eggs and modules.
195#   There will be a lot of gunk that was installed from packages;
196#   easy-install.pth in /usr/lib/ will tell you what was easy_installed.
197#   First use 'yum search' to see if the relevant package is now available
198#   as an RPM, and install that if it is.  If not, then use easy_install.
199#   Pass -Z to easy_install to install them unzipped, as some zipped eggs
200#   want to be able to write to ~/.python-eggs.  (Also makes sourcediving
201#   easier.)
202    cat /usr/lib/python2.6/site-packages/easy-install.pth
203# - Look at `gem list` for Ruby gems.
204#   Again, use 'yum search' and prefer RPMs, but failing that, 'gem install'.
205#       ezyang: rspec-rails depends on rspec, and will override the Yum
206#       package, so... don't use that RPM yet
207gem list --no-version > gem.txt
208    gem install $(gem list --no-version | grep -Fxvf - gem.txt)
209# - Look at `pear list` for Pear fruits (or whatever they're called).
210#   Yet again, 'yum search' for RPMs before resorting to 'pear install'.  Note
211#   that for things in the beta repo, you'll need 'pear install package-beta'.
212#   (you might get complaints about the php_scripts module; ignore them)
213pear list | tail -n +4 | cut -f 1 -d " " > pear.txt
214    pear config-set preferred_state beta
215    pear channel-update
216    pear install $(pear list | tail -n +4 | cut -f 1 -d " " | grep -Fxvf - pear.txt)
217# - Look at `pecl list` for PECL things.  'yum search', and if you must,
218#   'pecl install' needed items. If it doesn't work, try 'pear install
219#   pecl/foo' or 'pecl install foo-beta' or those two combined.
220pecl list | tail -n +4 | cut -f 1 -d " " > pecl.txt
221    pecl install --nodeps $(pecl list | tail -n +4 | cut -f 1 -d " " | grep -Fxvf - pecl.txt)
223# Setup some Python config
224    echo 'import site, os.path; site.addsitedir(os.path.expanduser("~/lib/python2.6/site-packages"))' > /usr/lib/python2.6/site-packages/00scripts-home.pth
226# Install the credentials.  There are a lot of things to remember here.
227# Be sure to make sure the permissions match up (ls -l on an existing
228# server!).
229scp root@$source_server:{/etc/{sql-mit-edu.cfg.php,daemon.keytab,pki/tls/private/scripts.key,signup-ldap-pw,whoisd-password},/home/logview/.k5login} .
230scp daemon.keytab signup-ldap-pw whoisd-password sql-mit-edu.cfg.php root@$server:/etc
231scp scripts.key root@$server:/etc/pki/tls/private
232scp .k5login root@$server:/home/logview
233    chown afsagent:afsagent /etc/daemon.keytab
234#   o The daemon.scripts keytab (will be daemon.scripts-test for test)
235    ls -l /etc/daemon.keytab
236#   o The SSL cert private key (real servers only)
237    ls -l /etc/pki/tls/private/scripts.key
238#   o The LDAP password for the signup process (real servers only)
239    ls -l /etc/signup-ldap-pw
240#   o The whoisd password (real servers only)
241    ls -l /etc/whoisd-password
242#   o Make sure logview's .k5login is correct (real servers only)
243    cat /home/logview/.k5login
245# Spin up OpenAFS.  This will fail if there's been a new kernel since
246# when you last tried.  In that case, you can hold on till later to
247# start OpenAFS.  This will take a little bit of time;
248    service openafs-client start
250# Check that fs sysname is correct.  You should see, among others,
251# 'amd64_fedoraX_scripts' (vary X) and 'scripts'. If it's not, you
252# probably did a distro upgrade and should update /etc/sysconfig/openafs.
253    fs sysname
255# [TEST SERVER] If you are setting up a test server, pay attention to
256# /etc/sysconfig/network-scripts and do not bind scripts' IP address.
257# You will also need to modify:
258#   o /etc/ldap.conf
259#       add: host
260#   o /etc/nss-ldapd.conf
261#       replace: uri *****
262#       with: uri ldap://
263#   o /etc/openldap/ldap.conf
264#       add: URI ldap://
265#            BASE dc=scripts,dc=mit,dc=edu
266#   o /etc/httpd/conf.d/vhost_ldap.conf
267#       replace: VhostLDAPUrl ****
268#       with: VhostLDAPUrl "ldap://,dc=scripts,dc=mit,dc=edu"
269#   o /etc/postfix/virtual-alias-{domains,maps}
270#       replace: server_host *****
271#       with: server_host = ldap://
272# to use instead of localhost.
273# XXX: someone should write sed scripts to do this
275# [TEST SERVER] If you are setting up a test server, afsagent's cronjob
276# will attempt to be renewing with the wrong credentials
277# (daemon.scripts). Change this:
278    vim /home/afsagent/renew # replace all mentions of
280# Set up replication (see ./install-ldap).
281# You'll need the LDAP keytab for this server: be sure to chown it
282# fedora-ds after you create the fedora-ds user
283    ls -l /etc/dirsrv/keytab
284    cat install-ldap
286# Make the services dirsrv, nslcd, nscd, postfix, and httpd start at
287# boot. Run chkconfig to make sure the set of services to be run is
288# correct.
289    chkconfig dirsrv on
290    chkconfig nslcd on
291    chkconfig nscd on
292    chkconfig postfix on
293    chkconfig httpd on
295# Check sql user credentials (needs to be done after LDAP is setup)
296    chown sql /etc/sql-mit-edu.cfg.php
298# Postfix doesn't actually deliver mail; fix this
299    cd /etc/postfix
300    postmap virtual
302# Munin might not be monitoring packages that were installed after it
303    munin-node-configure --suggest --shell | sh
305# Run fmtutil-sys --all, which does something that makes TeX work.
306# (Note: this errors on XeTeX which is ok.)
307    fmtutil-sys --all
309# Ensure that PHP isn't broken:
310    mkdir /tmp/sessions
311    chmod 01777 /tmp/sessions
313# Ensure fcgid isn't broken (should be 755)
314    ls -l /var/run | grep mod_fcgid
316# Fix etc by making sure none of our config files got overwritten
317    cd /etc
318    svn status -q
319    # Some usual candidates for clobbering include nsswitch.conf and
320    # sysconfig/openafs
322# ThisCell got clobbered, replace it with
323    echo "" > /usr/vice/etc/ThisCell
325# Reboot the machine to restore a consistent state, in case you
326# changed anything. (Note: Starting kdump fails (this is ok))
328# [OPTIONAL] Your machine's hostname is baked in at install time;
329# in the rare case you need to change it: it appears to be in:
330#   o /etc/sysconfig/network
331#   o your lvm thingies; probably don't need to edit
333# [TEST SERVER] More stuff for test servers
334#   - You need a self-signed SSL cert.  Generate with:
335    openssl req -new -x509 -keyout /etc/pki/tls/private/scripts.key -out /etc/pki/tls/certs/scripts.cert -nodes
336#     Also make /etc/pki/tls/certs/ca.pem match up
337#   - Make (/etc/aliases) root mail go to /dev/null, so we don't spam people
338#   - Edit /etc/httpd/conf.d/scripts-vhost-names.conf to have
339#     be an accepted vhost name
340#   - Look at the old test server and see what config changes are floating around
342# XXX: our SVN checkout should be updated to use
343# (repository and etc) once serving actually works.
344    cd /etc
345    svn switch --relocate svn://$source_server/ svn://
346    cd /usr/vice/etc
347    svn switch --relocate svn://$source_server/ svn://
348    cd /srv/repository
349    asbuild svn switch --relocate svn://$source_server/ svn://
350    asbuild svn up # verify works
Note: See TracBrowser for help on using the repository browser.