Différences

Ci-dessous, les différences entre deux révisions de la page.

--- smedev:install_sme_el6 [21/01/2013 10:46]
dani
+++ smedev:install_sme_el6 [25/01/2013 21:46] (Version actuelle)
dani
@@ Ligne 3: / Ligne 3: @@
 These are just my notes to get a EL6 system booting with all the e-smith and smeserver packages installed. As soon as I get it to boot, I'll put this info into contribs.org wiki.
-  * Start from a minimal el6 installation (I just installed a few utilities like htop, screen, rsync, vim, openssh-clients etc...)
+===== Minimal EL6 Install =====
-  * Disable graphical boot, it will be easier to see what's going on (and turn off selinux at kernel level, just to be sure)
+Just install a minimal el6 installation (I just installed a few utilities like htop, screen, rsync, vim, openssh-clients etc...). You can use either the DVD, the minimal CD install, a net install with PXE, it's up to you ;-)
+===== Disable graphical boot =====
+It will be easier to see what's going on (and turn off selinux at kernel level, just to be sure)
 <code bash>
 sed -i -e 's/rhgb quiet/selinux=0/g' /boot/grub/menu.lst
 </code>
-  * Disable SELinux:
+===== Disable SELinux =====
 <code bash>
 sed -i -e 's/SELINUX=.*/SELINUX=disabled/g' /etc/sysconfig/selinux
 </code>
-  * Remove selinux-policy-targeted and authconfig, they conflict with some e-smith/smeserver packages
+===== Remove selinux-policy-targeted and authconfig =====
+They conflict with some e-smith/smeserver packages
 <code bash>
 yum remove selinux-policy-targeted authconfig
 </code>
+===== Configure third party repo =====
   * Enable the EPEL repo
 <code bash>
@@ Ligne 30: / Ligne 40: @@
   * Configure SME repository (EL5 version for now)
 <code bash>
-cat<<'EOF' >> /etc/yum.repos.d/sme.repo
+cat<<'EOF' > /etc/yum.repos.d/sme.repo
 [smeaddons]
 enabled=1
@@ Ligne 97: / Ligne 107: @@
 yum install yum-plugin-priorities
 </code>
-  * Configure yum priorities
+===== Configure yum priorities =====
     * Set the base, updates and extras repo in /etc/yum.repos.d/CentOS-Base.repo a high priority (respectivly 50, 40, 50)
 <code bash>
@@ Ligne 165: / Ligne 176: @@
 priority=70
 </code>
-  * Now, a few packages have to be installed manually without dependancy check:
+===== Install som epackage without dependancy check =====
+A few packages have to be installed manually without dependancy check:
 <code bash>
 rpm -Uvh --nodeps http://sme-mirror.firewall-services.com/releases/8/smeupdates/x86_64/RPMS/smeserver-yum-2.2.0-20.el5.sme.noarch.rpm
@@ Ligne 175: / Ligne 188: @@
 yum install perl-Razor-Agent perl-Data-UUID --disablerepo=sme\* --disablerepo=epel
 </code>
-  * Now, you should be ready to install e-smith-base. For me, this pulls 222 packages (166MB). For some reason, the GPGKey is not recognized, so just run yum woith --nogpgcheck
+===== Install e-smith-base =====
+Now, you should be ready to install e-smith-base. For me, this pulls 222 packages (166MB). For some reason, the GPGKey is not recognized, so just run yum with --nogpgcheck
 <code bash>
 yum --nogpgcheck install e-smith-base
 </code>
-  * Now install all the remaining e-smith/smeserver packages
+===== Install the remaining e-smith/smeserver packages =====
+Now install all the remaining e-smith/smeserver packages
 <code bash>
 yum --nogpgcheck install e-smith\* smeserver\*
@@ Ligne 219: / Ligne 236: @@
 </code>
-  * Now install samba-client (e-smith-samba has been installed manually without dep check, otherwise it would have pulled samba3x package instead of samba)
+===== Install samba =====
+Now install samba-client (e-smith-samba has been installed manually without dep check, otherwise it would have pulled samba3x package instead of samba)
 <code bash>
 yum install samba-client
 </code>
-  * We've almost finished, we just have to replace upstart with the good old SysVinit, because upstart doesn't support the custom runlevel 7 SME uses
+===== Replace upstart with SysVinit =====
+We've almost finished, we just have to replace upstart with the good old SysVinit, because upstart doesn't support the custom runlevel 7 SME uses
 <code bash>
 rpm -e --nodeps upstart sysvinit-tools
@@ Ligne 230: / Ligne 250: @@
 </code>
-  * We need to copy some perl modules to a new directory, because @INC has changed in EL6:
+===== Move some perl modules =====
+We need to copy some perl modules to a new directory, because @INC has changed in EL6:
 <code bash>
 cp -a /usr/lib/perl5/site_perl/esmith/ /usr/share/perl5/vendor_perl/
 </code>
-  * We also need to create a new service entry in the DB, because the standard syslog package is now rsyslog:
+===== Create a service entry for rsyslog =====
+We need to create a new service entry in the DB, because the standard syslog package is now rsyslog:
 <code bash>
 /sbin/e-smith/db configuration set rsyslog service status enabled
@@ Ligne 241: / Ligne 265: @@
 </code>
-  * We can now try to post-upgrade
+===== Run post-upgrade =====
+We can now try to post-upgrade
 <code bash>
 /sbin/e-smith/signal-event post-upgrade
 </code>
-  * And reboot. Unfortunatly, I wasn't able to reboot properly at this stage, because of the upstart -> SysVinit transition (the error message is "/dev/initctl: No such file or directory") so I had to destroy the VM completly
+And reboot. Unfortunatly, I wasn't able to reboot properly at this stage, because of the upstart -> SysVinit transition (the error message is "/dev/initctl: No such file or directory") so I had to destroy the VM completly
+===== The system boots but completly broken :-) =====
 Ok, so the system should now boot, but is really not usable, after a quick look, at least the following doesn't work as expected:
   * The console on the first boot don't really work. It asks for the admin password (and BTW the password appears in cleartext) but it's all
-  * No network interfaces are detected by the console menu (because kudzu is not installed, only available in SME repo, and require an older python which clash the the one from EL6)
+  * No network interfaces are detected by the console menu (because kudzu is not installed, only available in SME repo, and require an older python which conflicts with the the one from EL6)
   * slapd won't start
   * httpd won't start
   * squid won't start
   * manually running expand-template has no effect (and no error message is printed). But signal-event seems to work (files get expanded and services restarted)
-  * as the network interface are not configured, all the services which requires the internel IP in their config (at least sshd squid and dnscache) won't start
+  * as the network interface are not configured, all the services which requires the internel IP (or the external one) in their config (at least sshd squid and dnscache) won't start
   * even if we manually add 0.0.0.0 after the ListenAddress line in /etc/ssh/sshd_config, we cannot connect using SSH (the daemon is running, but from a computer on the same network, I get a ssh_exchange_identification: Connection closed by remote host error)
   * qpsmtpd won't start (can't locate Qpsmtpd/TcpServer.pm in @INC)
   * There's probably a lot more which is not working
+  * [[what_is_not_working_on_el6|A more complete list of broken things]]