777
edits
Changes
Jump to navigation
Jump to search
Line 1,032:
Line 1,032: + + + + + + Line 1,114:
Line 1,120: − |- − |[https://src.koozali.org/smeserver/pwauth pwauth] − | − |X − |'''epel pwauth-2.3.10-25.el9.x86_64.rpm''' − but might need a patch Line 1,205:
Line 1,205: − + Line 1,254:
Line 1,254: − + + − + − + − install SME10 key+ Line 1,272:
Line 1,273: − + − + − dnf module enable mariadb:10.5 -y+ − + − dnf module enable python36:3.6 -y+ − + − + − a temp fix before smeserver-base-11.0.0-4.el8.sme − dnf install network-scripts rsyslog − et voilà, you have a non functional SME 11 over Rocky 8 minimal.+ + + − + + + − latter it will be good to also deactivate networkManager, to leave network service doing the job, but you also need to configure e-smith db for the server first.+ Line 1,292:
Line 1,295: + − + Line 1,303:
Line 1,307: − + + + + + + + + + + + + + + + + Line 1,312:
Line 1,331: − + − + − + + + + + + + + + + + + + + + + + + + + + + − +
→Bare install test
|builds ok
|builds ok
|https://koji.koozali.org/koji/buildinfo?buildID=258
|https://koji.koozali.org/koji/buildinfo?buildID=258
|-
|[https://src.koozali.org/smeserver/pwauth pwauth]
|jpp
|
|'''epel pwauth-2.3.10-25.el9.x86_64.rpm'''
but need a patch see https://bugs.koozali.org/show_bug.cgi?id=9442#c2 https://bugs.koozali.org/show_bug.cgi?id=9429#c3
|}
|}
|X
|X
|'''Rocky Linux BaseOS ppp-2.4.7-26.el8_1.x86_64.rpm'''
|'''Rocky Linux BaseOS ppp-2.4.7-26.el8_1.x86_64.rpm'''
|}
|}
update, and install those package, as the might get handy (nb iptraf is now iptraf-ng)
update, and install those package, as the might get handy (nb iptraf is now iptraf-ng)
dnf update -y
dnf update -y
dnf install mc htop iftop iptraf screen wget vim
dnf install mc htop iftop iptraf screen wget vim ppp traceroute setserial traceroute cyrus-sasl-md5 cyrus-sasl device-mapper-multipath dhcp-client fetchmail ftp iscsi-initiator-utils lm_sensors minicom mt-st sysfsutils tcpdump telnet wodim
* disable SELinux
* disable SELinux
gpgcheck=1
gpgcheck=1
enablegroups=1
enablegroups=1
[koji-sme11]
[koji-sme11-os]
enabled=1
enabled=1
name=koji-sme11
name=sme11-os
baseurl=<nowiki>http://koji.koozali.org/kojifiles/repos/dist-sme11-os-build/latest/x86_64/</nowiki>
baseurl=http://koji.koozali.org/kojifiles/repos/dist-sme11-build<nowiki/>/latest/x86_64/
gpgcheck=0
gpgcheck=0
enablegroups=1
enablegroups=1
install SME10 key
wget <nowiki>http://mirror.koozali.org/releases/10/smeos/x86_64/RPM-GPG-KEY-koozali</nowiki>
wget <nowiki>http://mirror.koozali.org/releases/10/smeos/x86_64/RPM-GPG-KEY-koozali</nowiki>
rpm --import RPM-GPG-KEY-koozali
rpm --import RPM-GPG-KEY-koozali
try installing SME11 packages, you might need to enable some repo using --enablerepo=smedev,epel,elrepo ... e.g.
try installing SME11 packages, you might need to enable some repo using --enablerepo=smedev,epel,elrepo ... e.g.
dnf module disable php:remi-8.2 -y
dnf module switch-to php:remi-8.3 -y
dnf module enable php:remi-8.1 -y
dnf module switch-to mariadb:10.5 -y
dnf module switch-to python36:3.6 -y
dnf module enable php:remi-8.1 -y
dnf module install php:remi-8.3
dnf install smeserver-* --enablerepo=koji-sme11-os,epel,remi,remi-safe,of --exclude=smeserver-mariadb* --exclude=smeserver-horde
dnf module install php:remi-8.1
{{Warning box|we will drop horde support, so we have no limitation in keeping php 8.1 which is obsolete, I suggest to go with 8.3 EOL 23 Nov 2026}}
dnf install smeserver-* --enablerepo=koji-sme11,epel,remi,remi-safe,of --exclude=smeserver-mariadb*
et voilà, you have a non functional SME 11 over Rocky 8 minimal.
Also remember that the legacy server manager (SM1) is not installed by the above, so, perhaps:
dnf install e-smith-manager
<nowiki>##</nowiki> this needs safer to be done directly on tty, not via ssh:
If you then find you get an SM1 login but the login just celars the fields and does not login, then you can fix this with:<syntaxhighlight lang="shell">
expand-template /etc/e-smith/web/common//cgi-bin/AuthTKT.cfg
</syntaxhighlight>'''<nowiki>##</nowiki> This is safer to be done directly on tty, not via ssh:'''
lget ride of Messup-Network-Manager NOW!
systemctl stop NetworkManager
systemctl stop NetworkManager
systemctl disable NetworkManager
systemctl disable NetworkManager
systemctl disable NetworkManager-wait-online.service
systemctl disable NetworkManager-wait-online.service
systemctl start networking
systemctl start networking
dnf remove NetworkManager* --noautoremove
then you could try, (you might have no network form there), unlink this to avoid root password locking
then you could try, (you might have no network form there), unlink this to avoid root password locking
unlink /etc/e-smith/events/post-install/S10init-passwords
unlink /etc/e-smith/events/post-install/S10init-passwords
signal-event post-install
/sbin/e-smith/signal-event post-install
if you want to avoid the reboot and configure loop for the moment you can do
if you want to avoid the reboot and configure loop for the moment you can do
passwd admin
passwd admin
/sbin/e-smith/db configuration setprop bootstrap-console Restore disabled
/sbin/e-smith/db configuration setprop bootstrap-console Restore disabled
then try to configure using console
then try to configure using console
console
/sbin/e-smith/console
To get better information from Server Manager 2 (this replaces the barfing dinosaur):<syntaxhighlight lang="shell">
mkdir -p /etc/e-smith/templates-custom/usr/share/smanager/conf/srvmngr.conf/
cat <<EOF > /etc/e-smith/templates-custom/usr/share/smanager/conf/srvmngr.conf/30Other
modules_dir => 'lib/SrvMngr/Controller',
webapp => 'smanager',
mode => 'development',
debug => 1,
\}
EOF
expand-template /usr/share/smanager/conf/srvmngr.conf
signal-event smanager-refresh
</syntaxhighlight>
then to be sure to get sshd back
then to be sure to get sshd back
/sbin/e-smith/db configuration setprop sshd status enabled PermitRootLogin yes access public
/sbin/e-smith/db configuration setprop sshd status enabled PermitRootLogin yes access public
# add PasswordAuthentication yes to enable password login
/sbin/e-smith/expand-template /etc/ssh/ssh_config
/sbin/e-smith/expand-template /etc/ssh/ssh_config
/sbin/e-smith/expand-template /etc/ssh/sshd_config
/sbin/e-smith/expand-template /etc/ssh/sshd_config
then you can
then you can
signal-event post-upgrade;
signal-event post-upgrade;
signal-event-reboot;
signal-event reboot;
or maybe
or maybe
signal-event bootstrap-console-save;
signal-event bootstrap-console-save;
signal-event-reboot
signal-event-reboot
<nowiki>##</nowiki> end of the part better on tty than on sshd
'''<nowiki>##</nowiki> end of the part better on tty than on sshd'''
you can also tidy a little
you can also tidy a little
dnf remove NetworkManager --noautoremove
dnf remove NetworkManager* --noautoremove
you might need avoid using sme* repo except for koji builsdsystem one as long as they are not populated to do so:
<syntaxhighlight lang="shell">
mkdir -p /etc/e-smith/templates-custom/etc/dnf/dnf.conf/
echo "#rem custom"> /etc/e-smith/templates-custom/etc/dnf/dnf.conf/20main_reposdir
expand-template /etc/dnf/dnf.conf
</syntaxhighlight>also you might want to plan ahead and put the koji repo for when you will have deleted the custom template.
vim /etc/yum.smerepos.d/koji.repo
and paste:
[koji-sme11-os]
enabled=1
name=sme11-os
baseurl=http://koji.koozali.org/kojifiles/repos/dist-sme11-build<nowiki/>/latest/x86_64/
gpgcheck=0
enablegroups=1
from there if you are lucky something went up and you can try to test , report as bug or event create a patch
from there if you are lucky it worked and you can try to test , report as bug or even create a patch
== Notes ==
== Notes ==