Definitely Not a Cloud
setup ClusterIt and check that it works. here’s a few hints.
check kernel options
dsh -e -g xen grep xen.gz /extlinux.conf dsh -e -g xen grep vmlinuz /extlinux.conf
what guests are running over the farm?
dsh -e -g xen "xl li | sed '1,2d' | awk '{print $1}'"
what is the current load?
dsh -e -g xen "xentop -b -i1"
full-blown new thin LV
TBD
new snapshot based on template’s LV
check what templates are available (empty origin)
lvs
e.g.
tpl=debian11jan2023
check what DRBD minors are available
grep minor /etc/drbd.d/*.res | sort -V -k4
e.g.
minor=1030
[optional] eventually tune the DRBD resource name
name=ltcpowa
create a new guest
dnc-new-resource.bash $tpl $minor $name
[optional] eventually tune the system hostname from the start
name=ltcpowa dnc-newguest-debian.bash $minor $name
check that the new guest is up and running
dnc-running-guest.bash $name
reach it remotely
ssh pmr.angrycow.ru -l root -p <PORT MAPPED AGAINST MINOR>
from an hypervisor as root
cd /data/guests/ for n in `seq 1 10`; do deployguest.bash netbsd.DATE.tar.gz renameguest.bash netbsd vadim$n xl create vadim$n/vadim$n done; unset n
from the mgmt node as user
update your remote connection settings,
grep dhcpd /var/log/syslog | tail sudo vi /etc/hosts ... vi ~/.ssh/config Host vadim* Port 2222 User root
make sure you have got the nobudget scripts available,
ll ~/bin/vadim*.ksh
to fix the hostnames with a for loop,
cd ~/bin/ for n in `seq 1 10`; do dsh -e -w vadim$n -s vadim.hostname.ksh vadim$n done; unset n
and eventually install the monit agent too
cd ~/bin/ vi vadim.monit.ksh #edit PASSWORD and MMONIT for n in `seq 1 10`; do dsh -e -w vadim$n -s vadim.monit.ksh vadim$n done; unset n