5 local total_attempts=$1; shift
8 echo "[NOTE] Waiting for cmd to return success: ${cmdstr}"
9 # shellcheck disable=SC2034
10 for attempt in $(seq "${total_attempts}"); do
11 # shellcheck disable=SC2015
12 eval "${cmdstr}" && break || true
13 echo -n '.'; sleep "${sleep_time}"
17 # Wait for MaaS commissioning/deploy to finish, retry on failure
18 function maas_fixup() {
19 local statuscmd="salt 'mas01*' --out yaml state.apply maas.machines.status"
20 wait_for 180 "${statuscmd} | tee /dev/stderr | " \
21 "grep -Eq '((Deployed|Ready): 5|status:Failed|status:Allocated)'"
22 # shellcheck disable=SC2155
23 local statusout=$(eval "${statuscmd}")
25 # shellcheck disable=SC2155
26 local fcnodes=$(echo "${statusout}" | \
27 grep -Po '(?<=system_id:)(.*)(?=,status:Failed commissioning)')
28 for node_system_id in ${fcnodes}; do
29 salt -C 'mas01*' state.apply maas.machines.delete \
30 pillar="{'system_id': '${node_system_id}'}"
32 if [ -n "${fcnodes}" ]; then
33 salt -C 'mas01*' state.apply maas.machines
37 # shellcheck disable=SC2155
38 local fdnodes=$(echo "${statusout}" | \
39 grep -Po '(?<=system_id:)(.*)(?=,status:(Failed deployment|Allocated))')
40 for node_system_id in ${fdnodes}; do
41 salt -C 'mas01*' state.apply maas.machines.mark_broken_fixed \
42 pillar="{'system_id': '${node_system_id}'}"
44 if [ -n "${fdnodes}" ]; then
45 salt -C 'mas01*' state.apply maas.machines.deploy
52 # MaaS rack/region controller, node commissioning
53 salt -C 'mas01*' cmd.run "add-apt-repository ppa:maas/stable"
55 salt -C 'mas01*' state.apply linux,salt,openssh,ntp
56 salt -C 'mas01*' state.apply linux.network.interface
57 salt -C 'mas01*' state.apply maas.pxe_nat
58 salt -C 'mas01*' state.apply maas.cluster
59 salt -C 'cfg01*' state.apply maas.pxe_route
61 wait_for 10 "salt -C 'mas01*' state.apply maas.region"
63 salt -C 'mas01*' state.apply maas.machines
64 wait_for 10 maas_fixup
66 # cleanup outdated salt keys
67 salt-key --out yaml | awk '!/^(minions|- cfg01|- mas01)/ {print $2}' | \
68 xargs -I{} salt-key -yd {}
70 # MaaS node deployment
71 salt -C 'mas01*' state.apply maas.machines.deploy
72 wait_for 10 maas_fixup
74 salt -C 'mas01*' pillar.item\
75 maas:region:admin:username \
76 maas:region:admin:password
78 # KVM, compute node prereqs (libvirt first), VCP deployment
79 salt -C '* and not cfg01* and not mas01*' saltutil.sync_all
81 salt -C 'kvm*' pkg.install bridge-utils
82 salt -C 'kvm*' state.apply linux.network
83 salt -C 'kvm*' system.reboot
84 wait_for 90 "! salt 'kvm*' test.ping | tee /dev/stderr | fgrep -q 'Not connected'"
86 salt -C '* and not cfg01* and not mas01*' state.apply linux,ntp
88 salt -C 'kvm*' state.sls libvirt
90 salt -C '* and not cfg01* and not mas01*' state.apply salt
91 salt -C 'kvm*' saltutil.sync_all
92 salt -C 'kvm*' state.sls salt.control
94 vcp_nodes=$(salt --out yaml 'kvm01*' pillar.get salt:control:cluster:internal:node | \
95 awk '/\s+\w+:$/ {gsub(/:$/, "*"); print $1}')
97 # Check all vcp nodes are available
99 while [ $rc -ne 0 ]; do
101 for node in $vcp_nodes; do
102 salt "$node" test.ping 2>/dev/null || { rc=$?; break; };
107 wait_for 10 "salt -C '* and not cfg01* and not mas01*' saltutil.sync_all"
108 wait_for 10 "salt -C 'E@^(?!cfg01|mas01|kvm|cmp00).*' state.apply salt"
109 wait_for 10 "! salt -C 'E@^(?!cfg01|mas01|kvm|cmp00).*' state.apply linux,ntp | " \
110 "tee /dev/stderr | fgrep -q 'Not connected'"
112 wait_for 10 "salt -C 'E@^(?!cfg01|mas01|kvm|cmp00).*' ssh.set_auth_key ${SUDO_USER} \
113 $(awk 'NR==1{print $2}' "$(eval echo "~${SUDO_USER}/.ssh/authorized_keys")")"
115 # Get the latest packages
116 salt '*' pkg.upgrade refresh=False