#!/bin/bash set -x function wait_for() { local total_attempts=$1; shift local cmdstr=$@ local sleep_time=10 echo "[NOTE] Waiting for cmd to return success: ${cmdstr}" for attempt in $(seq "${total_attempts}"); do eval "${cmdstr}" && break || true echo -n '.'; sleep "${sleep_time}" done } # Wait for MaaS commissioning/deploy to finish, retry on failure function maas_fixup() { local statuscmd="salt 'mas01*' --out yaml state.apply maas.machines.status" wait_for 180 "${statuscmd} | tee /dev/stderr | " \ "grep -Eq '((Deployed|Ready): 5|status:Failed|status:Allocated)'" local statusout=$(eval "${statuscmd}") local fcnodes=$(echo "${statusout}" | \ grep -Po '(?<=system_id:)(.*)(?=,status:Failed commissioning)') for node_system_id in ${fcnodes}; do salt -C 'mas01*' state.apply maas.machines.delete \ pillar="{'system_id': '${node_system_id}'}" done if [ -n "${fcnodes}" ]; then salt -C 'mas01*' state.apply maas.machines return 1 fi local fdnodes=$(echo "${statusout}" | \ grep -Po '(?<=system_id:)(.*)(?=,status:(Failed deployment|Allocated))') for node_system_id in ${fdnodes}; do salt -C 'mas01*' state.apply maas.machines.mark_broken_fixed \ pillar="{'system_id': '${node_system_id}'}" done if [ -n "${fdnodes}" ]; then salt -C 'mas01*' state.apply maas.machines.deploy return 1 fi return 0 } # MaaS rack/region controller, node commissioning salt -C 'mas01*' cmd.run "add-apt-repository ppa:maas/stable" salt -C 'mas01*' state.apply linux,salt,openssh,ntp salt -C 'mas01*' state.apply linux.network.interface salt -C 'mas01*' state.apply maas.pxe_nat salt -C 'mas01*' state.apply maas.cluster salt -C 'cfg01*' cmd.run \ "route add -net 192.168.11.0/24 gw ${MAAS_IP:-192.168.10.3}" wait_for 10 "salt -C 'mas01*' state.apply maas.region" salt -C 'mas01*' state.apply maas.machines wait_for 10 maas_fixup # cleanup outdated salt keys salt-key --out yaml | awk '!/^(minions|- cfg01|- mas01)/ {print $2}' | \ xargs -I{} salt-key -yd {} # MaaS node deployment salt -C 'mas01*' state.apply maas.machines.deploy wait_for 10 maas_fixup salt -C 'mas01*' pillar.item\ maas:region:admin:username \ maas:region:admin:password # KVM, compute node prereqs (libvirt first), VCP deployment salt -C '* and not cfg01* and not mas01*' saltutil.sync_all salt -C 'kvm*' pkg.install bridge-utils salt -C 'kvm*' state.apply linux.network salt -C 'kvm*' system.reboot wait_for 90 "! salt '*' test.ping | tee /dev/stderr | fgrep -q 'Not connected'" salt -C '* and not cfg01* and not mas01*' state.apply linux,ntp salt -C 'kvm*' state.sls libvirt salt -C '* and not cfg01* and not mas01*' state.apply salt salt -C 'kvm*' saltutil.sync_all salt -C 'kvm*' state.sls salt.control vcp_nodes=$(salt --out yaml 'kvm01*' pillar.get salt:control:cluster:internal:node | \ awk '/\s+\w+:$/ {gsub(/:$/, "*"); print $1}') # Check all vcp nodes are available rc=1 while [ $rc -ne 0 ]; do rc=0 for node in $vcp_nodes; do salt "$node" test.ping 2>/dev/null || { rc=$?; break; }; done sleep 5 done wait_for 10 "salt -C '* and not cfg01* and not mas01*' ssh.set_auth_key ${SUDO_USER} \ $(awk 'NR==1{print $2}' $(eval echo ~${SUDO_USER}/.ssh/authorized_keys))" wait_for 10 "salt -C '* and not cfg01* and not mas01*' saltutil.sync_all" wait_for 10 "salt -C '* and not cfg01* and not mas01*' state.apply salt" wait_for 10 "salt -C '* and not cfg01* and not mas01*' state.apply linux,ntp"