Cloudatcost : Différence entre versions
(Page créée avec « voici comment installer sme server 9 sur un vps chez cloudatcost.com installer centos 6 64 bits activer et configurer le ssh, ajoutez vos clefs ssh dans /root/.ssh/aut... ») |
|||
(8 révisions intermédiaires par le même utilisateur non affichées) | |||
Ligne 10 : | Ligne 10 : | ||
sed -i -e 's/rhgb quiet/selinux=0/g' /boot/grub/grub.conf | sed -i -e 's/rhgb quiet/selinux=0/g' /boot/grub/grub.conf | ||
sed -i -e 's/SELINUX=.*/SELINUX=disabled/g' /etc/selinux/config | sed -i -e 's/SELINUX=.*/SELINUX=disabled/g' /etc/selinux/config | ||
− | yum remove selinux-policy-targeted | + | yum remove selinux-policy-targeted -y |
+ | echo 0 > /selinux/enforce | ||
− | yum install wget | + | yum install wget -y |
mkdir /tmp/repo.bak | mkdir /tmp/repo.bak | ||
Ligne 29 : | Ligne 30 : | ||
yum --enablerepo=* clean all | yum --enablerepo=* clean all | ||
+ | |||
+ | #remove uneeded services | ||
+ | yum remove postfix -y | ||
Installez les paquets SME: | Installez les paquets SME: | ||
− | yum --disablerepo=* --enablerepo= | + | yum --disablerepo=* --enablerepo=smeupdates9,smeos9 --disablerepo=updates,base,smeupdates-testing9 install e-smith\* smeserver\* at bc bind-libs bind-utils cronie cronie-anacron crontabs cyrus-sasl-md5 dhcp ed eject elfutils elfutils-libs fetchmail ftp gettext glib hesiod iptraf iptstate isdn4k-utils libselinux-python libsysfs libxml2-python lm_sensors lockdev lsof lvm2 man minicom mkbootdisk mlocate mod_ssl mtools mt-st mutt net-snmp net-snmp-utils openldap-servers patch pcmciautils postgresql-libs pyxf86config rpm-build rp-pppoe setserial strace sudo sysfsutils syslinux tcpdump telnet time tmpwatch traceroute unzip usbutils usermode vim-common vim-enhanced wget wodim xz zip zlib php-pear-Net-URL perl-Compress-Raw-Bzip2 perl-IO-Compress-Bzip2 rsync tnef freeradius-ldap |
− | yum | + | yum install gettext --enablerepo=base,updates -y |
− | signal-event post- | + | you need the sme version not the centos one |
− | signal-event post-install | + | |
+ | yum update initscripts --disablerepo=* --enablerepo=smeos9,smeupdates9 | ||
+ | sometimes you need to downgrade instead of upgrade | ||
+ | |||
+ | update, but be sure your initscrips remains the sme one !!!! | ||
+ | |||
+ | yum upgrade --disablerepo=smeupdates-testing9 --enablerepo=base,updates --exclude=initscripts | ||
+ | |||
+ | /sbin/e-smith/signal-event post-upgrade | ||
+ | juste au cas ou on ne veut pas etre bloqué sans mot de passe | ||
+ | unlink /etc/e-smith/events/post-install/S10init-passwords | ||
+ | /sbin/e-smith/signal-event post-install | ||
assurez vous de garder un access ssh :D | assurez vous de garder un access ssh :D | ||
− | + | /sbin/e-smith/db configuration setprop sshd status enabled PermitRootLogin yes access public | |
− | signal-event remoteaccess-update | + | /sbin/e-smith/signal-event remoteaccess-update |
+ | |||
+ | config setprop httpd-admin ValidFrom YOURIP/255.255.255.255 | ||
+ | |||
+ | test you can open another sshd connexion after adding your ssh keys | ||
+ | |||
+ | si les services ne demarrent pas tous, il se peut que le fichier control : /service/*/supervise/control soient des fichiers regulier au lieu d'etre un fifo. | ||
+ | # rm /service/*/supervise/control | ||
+ | rm: remove regular file `/service/dhcpd/supervise/control'? y | ||
+ | rm: remove regular file `/service/ftp/supervise/control'? y | ||
+ | rm: remove regular file `/service/httpd-admin/supervise/control'? y | ||
+ | rm: remove regular file `/service/httpd-e-smith/supervise/control'? y | ||
+ | rm: remove regular file `/service/pptpd/supervise/control'? y | ||
+ | rm: remove regular file `/service/radiusd/supervise/control'? y | ||
+ | rm: remove regular file `/service/sshd/supervise/control'? y | ||
+ | |||
+ | devrait aider. dites oui a tous les fichiers reguliers, mais n'effacez pas les fifo. (au pire il seront recrés au reboot) | ||
+ | |||
+ | #configurez votre serveur | ||
+ | console | ||
+ | |||
+ | # forcer les quotas check | ||
+ | touch /forcequotacheck | ||
+ | |||
+ | # because of disk being slow you need to increase timeout; and make it permanent on reboot | ||
+ | echo 360 > /sys/block/sda/device/timeout | ||
+ | echo 'cat /sys/block/sda/device/timeout >/tmp/timeout | ||
+ | echo 360 > /sys/block/sda/device/timeout' >> /etc/rc.d/rc.local | ||
+ | # desactiver la restauration de backup | ||
+ | # si vous preferez regler votre mot de passe sans apsser par le bootstrap console ( l'interface de cloudatcost est en http pas en https) | ||
+ | config set PasswordSet yes | ||
+ | |||
+ | rebootez: | ||
+ | /sbin/e-smith/signal-event post-upgrade ;/sbin/e-smith/signal-event reboot | ||
− | + | il se peut que runsvdir fasse une erreur sur un fichier regulier pour wan a cette étape au lieu d'un fifo: | |
− | + | ||
+ | rm /service/wan/supervise/control |
Version actuelle en date du 28 avril 2017 à 13:21
voici comment installer sme server 9 sur un vps chez cloudatcost.com
installer centos 6 64 bits
activer et configurer le ssh, ajoutez vos clefs ssh dans /root/.ssh/authorized_keys
enlever selinux
sed -i -e 's/rhgb quiet/selinux=0/g' /boot/grub/grub.conf sed -i -e 's/SELINUX=.*/SELINUX=disabled/g' /etc/selinux/config yum remove selinux-policy-targeted -y echo 0 > /selinux/enforce
yum install wget -y
mkdir /tmp/repo.bak mv /etc/yum.repos.d/*.repo /tmp/repo.bak/
ajouter le contenu de /etc/yum.smerepo/sme.repo d'une sme9:
wget -O /etc/yum.repos.d/sme9.repo http://bugs.contribs.org/attachment.cgi?id=3652
Import needed GPG Keys
rpm --import http://sme-mirror.firewall-services.com/releases/8/smeos/x86_64/RPM-GPG-KEY-SMEServer rpm --import http://apt.sw.be/RPM-GPG-KEY.dag.txt rpm --import https://fedoraproject.org/static/217521F6.txt
Clean yum cache
yum --enablerepo=* clean all
- remove uneeded services
yum remove postfix -y
Installez les paquets SME:
yum --disablerepo=* --enablerepo=smeupdates9,smeos9 --disablerepo=updates,base,smeupdates-testing9 install e-smith\* smeserver\* at bc bind-libs bind-utils cronie cronie-anacron crontabs cyrus-sasl-md5 dhcp ed eject elfutils elfutils-libs fetchmail ftp gettext glib hesiod iptraf iptstate isdn4k-utils libselinux-python libsysfs libxml2-python lm_sensors lockdev lsof lvm2 man minicom mkbootdisk mlocate mod_ssl mtools mt-st mutt net-snmp net-snmp-utils openldap-servers patch pcmciautils postgresql-libs pyxf86config rpm-build rp-pppoe setserial strace sudo sysfsutils syslinux tcpdump telnet time tmpwatch traceroute unzip usbutils usermode vim-common vim-enhanced wget wodim xz zip zlib php-pear-Net-URL perl-Compress-Raw-Bzip2 perl-IO-Compress-Bzip2 rsync tnef freeradius-ldap
yum install gettext --enablerepo=base,updates -y
you need the sme version not the centos one
yum update initscripts --disablerepo=* --enablerepo=smeos9,smeupdates9
sometimes you need to downgrade instead of upgrade
update, but be sure your initscrips remains the sme one !!!!
yum upgrade --disablerepo=smeupdates-testing9 --enablerepo=base,updates --exclude=initscripts
/sbin/e-smith/signal-event post-upgrade
juste au cas ou on ne veut pas etre bloqué sans mot de passe
unlink /etc/e-smith/events/post-install/S10init-passwords /sbin/e-smith/signal-event post-install
assurez vous de garder un access ssh :D
/sbin/e-smith/db configuration setprop sshd status enabled PermitRootLogin yes access public /sbin/e-smith/signal-event remoteaccess-update
config setprop httpd-admin ValidFrom YOURIP/255.255.255.255
test you can open another sshd connexion after adding your ssh keys
si les services ne demarrent pas tous, il se peut que le fichier control : /service/*/supervise/control soient des fichiers regulier au lieu d'etre un fifo.
# rm /service/*/supervise/control rm: remove regular file `/service/dhcpd/supervise/control'? y rm: remove regular file `/service/ftp/supervise/control'? y rm: remove regular file `/service/httpd-admin/supervise/control'? y rm: remove regular file `/service/httpd-e-smith/supervise/control'? y rm: remove regular file `/service/pptpd/supervise/control'? y rm: remove regular file `/service/radiusd/supervise/control'? y rm: remove regular file `/service/sshd/supervise/control'? y
devrait aider. dites oui a tous les fichiers reguliers, mais n'effacez pas les fifo. (au pire il seront recrés au reboot)
- configurez votre serveur
console
- forcer les quotas check
touch /forcequotacheck
- because of disk being slow you need to increase timeout; and make it permanent on reboot
echo 360 > /sys/block/sda/device/timeout echo 'cat /sys/block/sda/device/timeout >/tmp/timeout echo 360 > /sys/block/sda/device/timeout' >> /etc/rc.d/rc.local
- desactiver la restauration de backup
- si vous preferez regler votre mot de passe sans apsser par le bootstrap console ( l'interface de cloudatcost est en http pas en https)
config set PasswordSet yes
rebootez:
/sbin/e-smith/signal-event post-upgrade ;/sbin/e-smith/signal-event reboot
il se peut que runsvdir fasse une erreur sur un fichier regulier pour wan a cette étape au lieu d'un fifo:
rm /service/wan/supervise/control