1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
|
#!/bin/bash -e
# shellcheck disable=SC2155,SC1001,SC2015,SC2128
##############################################################################
# Copyright (c) 2018 Mirantis Inc., Enea AB and others.
# All rights reserved. This program and the accompanying materials
# are made available under the terms of the Apache License, Version 2.0
# which accompanies this distribution, and is available at
# http://www.apache.org/licenses/LICENSE-2.0
##############################################################################
#
# Library of shell functions used by deploy script on jumpserver:
# - base cloud image (used by FN VMs and VCP VMs) processing:
# * download;
# * tooling for offline image modification (without libguestfs);
# * package pre-installation (requires nbd, loop krn mods);
# - virtualized hosts processing:
# * virsh-managed VMs boilerplate;
# * salt master container tooling;
# * virsh & docker network plumbing;
# etc.
##############################################################################
# private helper functions
##############################################################################
function __get_base_image {
local base_image=$1
local image_dir=$2
mkdir -p "${image_dir}"
wget --progress=dot:giga -P "${image_dir}" -N "${base_image}"
}
function __kernel_modules {
# Load mandatory kernel modules: loop, nbd
local image_dir=$1
test -e /dev/loop-control || sudo modprobe loop
if sudo modprobe nbd max_part=8 || sudo modprobe -f nbd max_part=8; then
return 0
fi
if [ -e /dev/nbd0 ]; then return 0; fi # nbd might be inbuilt
# CentOS (or RHEL family in general) do not provide 'nbd' out of the box
echo "[WARN] 'nbd' kernel module cannot be loaded!"
if [ ! -e /etc/redhat-release ]; then
echo "[ERROR] Non-RHEL system detected, aborting!"
echo "[ERROR] Try building 'nbd' manually or install it from a 3rd party."
exit 1
fi
# Best-effort attempt at building a non-maintaned kernel module
local __baseurl='http://vault.centos.org/centos'
local __subdir='Source/SPackages'
local __uname_r=$(uname -r)
local __uname_m=$(uname -m)
if [ "${__uname_m}" = 'x86_64' ]; then
__srpm="kernel-${__uname_r%.${__uname_m}}.src.rpm"
else
# NOTE: fmt varies across releases (e.g. kernel-alt-4.11.0-44.el7a.src.rpm)
__srpm="kernel-alt-${__uname_r%.${__uname_m}}.src.rpm"
fi
local __found='n'
local __versions=$(curl -s "${__baseurl}/" | grep -Po 'href="\K7\.[\d\.]+')
for ver in ${__versions}; do
for comp in os updates; do
local url="${__baseurl}/${ver}/${comp}/${__subdir}/${__srpm}"
if wget "${url}" -O "${image_dir}/${__srpm}" > /dev/null 2>&1; then
__found='y'; break 2
fi
done
done
if [ "${__found}" = 'n' ]; then
echo "[ERROR] Can't find the linux kernel SRPM for: ${__uname_r}"
echo "[ERROR] 'nbd' module cannot be built, aborting!"
echo "[ERROR] Try 'yum upgrade' or building 'nbd' krn module manually ..."
exit 1
fi
rpm -ivh "${image_dir}/${__srpm}" 2> /dev/null
mkdir -p ~/rpmbuild/{BUILD,BUILDROOT,RPMS,SOURCES,SPECS,SRPMS}
# shellcheck disable=SC2016
echo '%_topdir %(echo $HOME)/rpmbuild' > ~/.rpmmacros
(
cd ~/rpmbuild/SPECS
rpmbuild -bp --nodeps --target="${__uname_m}" kernel*.spec
cd ~/rpmbuild/BUILD/"${__srpm%.src.rpm}"/linux-*
sed -i 's/^.*\(CONFIG_BLK_DEV_NBD\).*$/\1=m/g' .config
# http://centosfaq.org/centos/nbd-does-not-compile-for-3100-514262el7x86_64
if grep -Rq 'REQ_TYPE_DRV_PRIV' drivers/block; then
sed -i 's/REQ_TYPE_SPECIAL/REQ_TYPE_DRV_PRIV/g' drivers/block/nbd.c
fi
gunzip -c "/boot/symvers-${__uname_r}.gz" > Module.symvers
make prepare modules_prepare
make M=drivers/block -j
modinfo drivers/block/nbd.ko
sudo mkdir -p "/lib/modules/${__uname_r}/extra/"
sudo cp drivers/block/nbd.ko "/lib/modules/${__uname_r}/extra/"
)
sudo depmod -a
sudo modprobe nbd max_part=8 || sudo modprobe -f nbd max_part=8
}
function __mount_image {
local image=$1
local image_dir=$2
OPNFV_MNT_DIR="${image_dir}/ubuntu"
# Find free nbd, loop devices
for dev in '/sys/class/block/nbd'*; do
if [ "$(cat "${dev}/size")" = '0' ]; then
OPNFV_NBD_DEV=/dev/$(basename "${dev}")
break
fi
done
OPNFV_LOOP_DEV=$(sudo losetup -f)
OPNFV_MAP_DEV=/dev/mapper/$(basename "${OPNFV_NBD_DEV}")p1
export OPNFV_MNT_DIR OPNFV_LOOP_DEV
[ -n "${OPNFV_NBD_DEV}" ] && [ -n "${OPNFV_LOOP_DEV}" ] || exit 1
qemu-img resize "${image_dir}/${image}" 3G
sudo qemu-nbd --connect="${OPNFV_NBD_DEV}" --aio=native --cache=none \
"${image_dir}/${image}"
sudo kpartx -av "${OPNFV_NBD_DEV}"
# Hardcode partition index to 1, unlikely to change for Ubuntu UCA image
sudo partx -uvn 1:1 "${OPNFV_NBD_DEV}"
if sudo growpart "${OPNFV_NBD_DEV}" 1; then
sudo kpartx -u "${OPNFV_NBD_DEV}"
sudo e2fsck -pf "${OPNFV_MAP_DEV}"
sudo resize2fs "${OPNFV_MAP_DEV}"
else
sleep 5 # /dev/nbdNp1 takes some time to come up
fi
sudo partx -d "${OPNFV_NBD_DEV}"
# grub-update does not like /dev/nbd*, so use a loop device to work around it
sudo losetup "${OPNFV_LOOP_DEV}" "${OPNFV_MAP_DEV}"
mkdir -p "${OPNFV_MNT_DIR}"
sudo mount "${OPNFV_LOOP_DEV}" "${OPNFV_MNT_DIR}"
sudo mount -t proc proc "${OPNFV_MNT_DIR}/proc"
sudo mount -t sysfs sys "${OPNFV_MNT_DIR}/sys"
sudo mount -o bind /dev "${OPNFV_MNT_DIR}/dev"
sudo mkdir -p "${OPNFV_MNT_DIR}/run/resolvconf"
sudo cp /etc/resolv.conf "${OPNFV_MNT_DIR}/run/resolvconf"
echo "GRUB_DISABLE_OS_PROBER=true" | \
sudo tee -a "${OPNFV_MNT_DIR}/etc/default/grub"
sudo sed -i -e 's/^\(GRUB_TIMEOUT\)=.*$/\1=1/g' -e 's/^GRUB_HIDDEN.*$//g' \
"${OPNFV_MNT_DIR}/etc/default/grub"
}
function __apt_repos_pkgs_image {
local apt_key_urls=(${1//,/ })
local all_repos=(${2//,/ })
local pkgs_i=(${3//,/ })
local pkgs_r=(${4//,/ })
[ -n "${OPNFV_MNT_DIR}" ] || exit 1
# APT keys
if [ "${#apt_key_urls[@]}" -gt 0 ]; then
for apt_key in "${apt_key_urls[@]}"; do
sudo chroot "${OPNFV_MNT_DIR}" /bin/bash -c \
"wget -qO - '${apt_key}' | apt-key add -"
done
fi
# Additional repositories
for repo_line in "${all_repos[@]}"; do
# <repo_name>|<repo prio>|deb|[arch=<arch>]|<repo url>|<dist>|<repo comp>
local repo=(${repo_line//|/ })
[ "${#repo[@]}" -gt 5 ] || continue
# NOTE: Names and formatting are compatible with Salt linux.system.repo
cat <<-EOF | sudo tee "${OPNFV_MNT_DIR}/etc/apt/preferences.d/${repo[0]}"
Package: *
Pin: release a=${repo[-2]}
Pin-Priority: ${repo[1]}
EOF
echo "${repo[@]:2}" | sudo tee \
"${OPNFV_MNT_DIR}/etc/apt/sources.list.d/${repo[0]}.list"
done
# Install packages
if [ "${#pkgs_i[@]}" -gt 0 ]; then
sudo DEBIAN_FRONTEND="noninteractive" \
chroot "${OPNFV_MNT_DIR}" apt-get update
sudo DEBIAN_FRONTEND="noninteractive" FLASH_KERNEL_SKIP="true" \
chroot "${OPNFV_MNT_DIR}" apt-get install -y "${pkgs_i[@]}"
fi
# Remove packages
if [ "${#pkgs_r[@]}" -gt 0 ]; then
sudo DEBIAN_FRONTEND="noninteractive" FLASH_KERNEL_SKIP="true" \
chroot "${OPNFV_MNT_DIR}" apt-get purge -y "${pkgs_r[@]}"
fi
# Disable cloud-init metadata service datasource
sudo mkdir -p "${OPNFV_MNT_DIR}/etc/cloud/cloud.cfg.d"
echo "datasource_list: [ NoCloud, None ]" | sudo tee \
"${OPNFV_MNT_DIR}/etc/cloud/cloud.cfg.d/95_real_datasources.cfg"
}
function __cleanup_vms {
# clean up existing nodes
for node in $(${VIRSH} list --name | grep -P '\w{3}\d{2}'); do
${VIRSH} destroy "${node}" 2>/dev/null || true
done
for node in $(${VIRSH} list --name --all | grep -P '\w{3}\d{2}'); do
${VIRSH} domblklist "${node}" | awk '/^.da/ {print $2}' | \
xargs --no-run-if-empty -I{} sudo rm -f {}
${VIRSH} undefine "${node}" --remove-all-storage --nvram || \
${VIRSH} undefine "${node}" --remove-all-storage
done
}
##############################################################################
# public functions
##############################################################################
function prepare_vms {
local base_image=$1; shift
local image_dir=$1; shift
local repos_pkgs_str=$1; shift # ^-sep list of repos, pkgs to install/rm
local image=base_image_opnfv_fuel.img
local vcp_image=${image%.*}_vcp.img
local _o=${base_image/*\/}
[ -n "${image_dir}" ] || exit 1
cleanup_uefi
__cleanup_vms
__get_base_image "${base_image}" "${image_dir}"
IFS='^' read -r -a repos_pkgs <<< "${repos_pkgs_str}"
local _h=$(echo "${repos_pkgs_str}.$(md5sum "${image_dir}/${_o}")" | \
md5sum | cut -c -8)
local _tmp="${image%.*}.${_h}.img"
echo "[INFO] Lookup cache / build patched base image for fingerprint: ${_h}"
if [ "${image_dir}/${_tmp}" -ef "${image_dir}/${image}" ]; then
echo "[INFO] Patched base image found"
else
# shellcheck disable=SC2115
rm -rf "${image_dir}/${image%.*}"*
if [[ ! "${repos_pkgs_str}" =~ ^\^+$ ]]; then
echo "[INFO] Patching base image ..."
cp "${image_dir}/${_o}" "${image_dir}/${_tmp}"
__kernel_modules "${image_dir}"
__mount_image "${_tmp}" "${image_dir}"
__apt_repos_pkgs_image "${repos_pkgs[@]:0:4}"
cleanup_mounts
else
echo "[INFO] No patching required, using vanilla base image"
ln -sf "${image_dir}/${_o}" "${image_dir}/${_tmp}"
fi
ln -sf "${image_dir}/${_tmp}" "${image_dir}/${image}"
fi
# VCP VMs base image specific changes
if [[ ! "${repos_pkgs_str}" =~ \^{3}$ ]] && [ -n "${repos_pkgs[*]:4}" ]; then
echo "[INFO] Lookup cache / build patched VCP image for md5sum: ${_h}"
_tmp="${vcp_image%.*}.${_h}.img"
if [ "${image_dir}/${_tmp}" -ef "${image_dir}/${vcp_image}" ]; then
echo "[INFO] Patched VCP image found"
else
echo "[INFO] Patching VCP image ..."
cp "${image_dir}/${image}" "${image_dir}/${_tmp}"
__kernel_modules "${image_dir}"
__mount_image "${_tmp}" "${image_dir}"
__apt_repos_pkgs_image "${repos_pkgs[@]:4:4}"
cleanup_mounts
ln -sf "${image_dir}/${_tmp}" "${image_dir}/${vcp_image}"
fi
fi
}
function create_networks {
local all_vnode_networks=("$@")
# create required networks
for net in "mcpcontrol" "${all_vnode_networks[@]}"; do
if ${VIRSH} net-info "${net}" >/dev/null 2>&1; then
${VIRSH} net-destroy "${net}" || true
${VIRSH} net-undefine "${net}"
fi
# in case of custom network, host should already have the bridge in place
if [ -f "virsh_net/net_${net}.xml" ] && \
[ ! -d "/sys/class/net/${net}/bridge" ]; then
${VIRSH} net-define "virsh_net/net_${net}.xml"
${VIRSH} net-autostart "${net}"
${VIRSH} net-start "${net}"
fi
done
sudo ip link del veth_mcp0 || true
sudo ip link del veth_mcp2 || true
# Create systemd service for veth creation after reboot
FUEL_VETHC_SERVICE="/etc/systemd/system/opnfv-fuel-vethc.service"
FUEL_VETHA_SERVICE="/etc/systemd/system/opnfv-fuel-vetha.service"
test -f /usr/sbin/ip && PREFIX=/usr/sbin || PREFIX=/sbin
cat <<-EOF | sudo tee "${FUEL_VETHC_SERVICE}"
[Unit]
After=libvirtd.service
Before=docker.service
[Service]
ExecStart=/bin/sh -ec '\
${PREFIX}/ip link add veth_mcp0 type veth peer name veth_mcp1 && \
${PREFIX}/ip link add veth_mcp2 type veth peer name veth_mcp3 && \
${PREFIX}/ip link set veth_mcp0 up mtu 9000 && \
${PREFIX}/ip link set veth_mcp1 up mtu 9000 && \
${PREFIX}/ip link set veth_mcp2 up mtu 9000 && \
${PREFIX}/ip link set veth_mcp3 up mtu 9000'
EOF
cat <<-EOF | sudo tee "${FUEL_VETHA_SERVICE}"
[Unit]
StartLimitInterval=200
StartLimitBurst=10
After=opnfv-fuel-vethc.service
[Service]
Restart=on-failure
RestartSec=10
ExecStartPre=/bin/sh -ec '\
${PREFIX}/brctl showstp ${all_vnode_networks[0]} > /dev/null 2>&1 && \
${PREFIX}/brctl showstp ${all_vnode_networks[1]} > /dev/null 2>&1'
ExecStart=/bin/sh -ec '\
${PREFIX}/brctl addif ${all_vnode_networks[0]} veth_mcp0 && \
${PREFIX}/brctl addif ${all_vnode_networks[1]} veth_mcp2 && \
while ${PREFIX}/ip rule del to ${SALT_MASTER} iif docker0 table 200 2>/dev/null; do true; done && \
${PREFIX}/ip rule add to ${SALT_MASTER} iif docker0 table 200 && \
${PREFIX}/ip route replace ${SALT_MASTER} dev ${all_vnode_networks[0]} table 200'
EOF
sudo ln -sf "${FUEL_VETHC_SERVICE}" "/etc/systemd/system/multi-user.target.wants/"
sudo ln -sf "${FUEL_VETHA_SERVICE}" "/etc/systemd/system/multi-user.target.wants/"
sudo systemctl daemon-reload
sudo systemctl restart opnfv-fuel-vethc
sudo systemctl restart opnfv-fuel-vetha
}
function create_vms {
local image_dir=$1; shift
local image=base_image_opnfv_fuel.img
# vnode data should be serialized with the following format:
# <name0>,<disks0>,<ram0>,<vcpu0>[,<sockets0>,<cores0>,<threads0>[,<cell0name0>,<cell0memory0>,
# <cell0cpus0>,<cell1name0>,<cell1memory0>,<cell1cpus0>]]|<name1>,...'
IFS='|' read -r -a vnodes <<< "$1"; shift
# AArch64: prepare arch specific arguments
local virt_extra_args=""
if [ "$(uname -i)" = "aarch64" ]; then
# No Cirrus VGA on AArch64, use virtio instead
virt_extra_args="$virt_extra_args --video=virtio"
fi
# create vms with specified options
for serialized_vnode_data in "${vnodes[@]}"; do
if [ -z "${serialized_vnode_data}" ]; then continue; fi
IFS=',' read -r -a vnode_data <<< "${serialized_vnode_data}"
IFS=';' read -r -a disks_data <<< "${vnode_data[1]}"
# Create config ISO and resize OS disk image for each foundation node VM
./create-config-drive.sh -k "$(basename "${SSH_KEY}").pub" \
-u 'user-data.sh' -h "${vnode_data[0]}" "${image_dir}/mcp_${vnode_data[0]}.iso"
cp "${image_dir}/${image}" "${image_dir}/mcp_${vnode_data[0]}.qcow2"
qemu-img resize "${image_dir}/mcp_${vnode_data[0]}.qcow2" "${disks_data[0]}"
# Prepare additional drives if present
idx=0
virt_extra_storage=
for dsize in "${disks_data[@]:1}"; do
((idx+=1))
qcow_file="${image_dir}/mcp_${vnode_data[0]}_${idx}.qcow2"
qemu-img create "${qcow_file}" "${dsize}"
virt_extra_storage+=" --disk path=${qcow_file},format=qcow2,bus=virtio,cache=none,io=native"
done
# prepare VM CPU model, count, topology (optional), NUMA cells (optional, requires topo)
local virt_cpu_args=' --cpu host-passthrough'
local idx=7 # cell0.name index in serialized data
while [ -n "${vnode_data[${idx}]}" ]; do
virt_cpu_args+=",${vnode_data[${idx}]}.memory=${vnode_data[$((idx + 1))]}"
virt_cpu_args+=",${vnode_data[${idx}]}.cpus=${vnode_data[$((idx + 2))]}"
idx=$((idx + 3))
done
virt_cpu_args+=" --vcpus vcpus=${vnode_data[3]}"
if [ -n "${vnode_data[6]}" ]; then
virt_cpu_args+=",sockets=${vnode_data[4]},cores=${vnode_data[5]},threads=${vnode_data[6]}"
fi
# prepare network args
local vnode_networks=("$@")
local net_args=
for net in "${vnode_networks[@]}"; do
net_args="${net_args} --network bridge=${net},model=virtio"
done
[ ! -e "${image_dir}/virt-manager" ] || VIRT_PREFIX="${image_dir}/virt-manager/"
# shellcheck disable=SC2086
${VIRT_PREFIX}${VIRSH/virsh/virt-install} --name "${vnode_data[0]}" \
${virt_cpu_args} --accelerate \
${net_args} \
--ram "${vnode_data[2]}" \
--disk path="${image_dir}/mcp_${vnode_data[0]}.qcow2",format=qcow2,bus=virtio,cache=none,io=native \
${virt_extra_storage} \
--os-type linux --os-variant none \
--boot hd --vnc --console pty --autostart --noreboot \
--disk path="${image_dir}/mcp_${vnode_data[0]}.iso",device=cdrom \
--noautoconsole \
${virt_extra_args}
done
}
function reset_vms {
local vnodes=("$@")
local cmd_str="ssh ${SSH_OPTS} ${SSH_SALT}"
# reset non-infrastructure vms, wait for them to come back online
for node in "${vnodes[@]}"; do
${VIRSH} reset "${node}"
done
for node in "${vnodes[@]}"; do
wait_for 20.0 "${cmd_str} sudo salt -C '${node}*' saltutil.sync_all"
done
}
function start_vms {
local vnodes=("$@")
# start vms
for node in "${vnodes[@]}"; do
${VIRSH} start "${node}"
sleep $((RANDOM%5+1))
done
}
function prepare_containers {
local image_dir=$1
[ -n "${image_dir}" ] || exit 1
[ -n "${MCP_REPO_ROOT_PATH}" ] || exit 1
[ ! -e "${image_dir}/docker-compose" ] || COMPOSE_PREFIX="${image_dir}/"
"${COMPOSE_PREFIX}docker-compose" -f docker-compose/docker-compose.yaml down
if [[ ! "${MCP_DOCKER_TAG}" =~ 'verify' ]]; then
"${COMPOSE_PREFIX}docker-compose" -f docker-compose/docker-compose.yaml pull
fi
# overwrite hosts only on first container up, to preserve cluster nodes
sudo cp docker-compose/files/hosts "${image_dir}/hosts"
sudo rm -rf "${image_dir}/"{salt,pki,mas01/etc} "${image_dir}/nodes/"*
find "${image_dir}/mas01/var/lib/" \
-mindepth 2 -maxdepth 2 -not -name boot-resources \
-exec sudo rm -rf {} \; || true
mkdir -p "${image_dir}/"{salt/master.d,salt/minion.d}
if grep -q -e 'maas' 'docker-compose/docker-compose.yaml'; then
chmod +x docker-compose/files/entrypoint*.sh
# Apparmor workaround for bind9 inside Docker containers using AUFS
for profile in 'usr.sbin.ntpd' 'usr.sbin.named' \
'usr.sbin.dhcpd' 'usr.bin.tcpdump'; do
if [ -e "/etc/apparmor.d/${profile}" ] && \
[ ! -e "/etc/apparmor.d/disable/${profile}" ]; then
sudo ln -sf "/etc/apparmor.d/${profile}" "/etc/apparmor.d/disable/"
sudo apparmor_parser -R "/etc/apparmor.d/${profile}" || true
fi
done
fi
}
function start_containers {
local image_dir=$1
[ -n "${image_dir}" ] || exit 1
[ ! -e "${image_dir}/docker-compose" ] || COMPOSE_PREFIX="${image_dir}/"
"${COMPOSE_PREFIX}docker-compose" -f docker-compose/docker-compose.yaml up -d
}
function check_connection {
local total_attempts=60
local sleep_time=5
set +e
echo '[INFO] Attempting to get into Salt master ...'
# wait until ssh on Salt master is available
# shellcheck disable=SC2034
for attempt in $(seq "${total_attempts}"); do
# shellcheck disable=SC2086
ssh ${SSH_OPTS} "ubuntu@${SALT_MASTER}" uptime
case $? in
0) echo "${attempt}> Success"; break ;;
*) echo "${attempt}/${total_attempts}> ssh server ain't ready yet, waiting for ${sleep_time} seconds ..." ;;
esac
sleep $sleep_time
done
set -e
}
function cleanup_mounts {
# Remove any mounts, loop and/or nbd devs created while patching base image
if [ -n "${OPNFV_MNT_DIR}" ] && [ -d "${OPNFV_MNT_DIR}" ]; then
if [ -f "${OPNFV_MNT_DIR}/boot/grub/grub.cfg" ]; then
# Grub thinks it's running from a live CD
sudo sed -i -e 's/^\s*set root=.*$//g' -e 's/^\s*loopback.*$//g' \
"${OPNFV_MNT_DIR}/boot/grub/grub.cfg"
fi
sudo rm -f "${OPNFV_MNT_DIR}/run/resolvconf/resolv.conf"
sync
if mountpoint -q "${OPNFV_MNT_DIR}"; then
sudo umount -l "${OPNFV_MNT_DIR}" || true
fi
fi
if [ -n "${OPNFV_LOOP_DEV}" ] && \
sudo losetup "${OPNFV_LOOP_DEV}" 1>&2 > /dev/null; then
sudo losetup -d "${OPNFV_LOOP_DEV}"
fi
if [ -n "${OPNFV_NBD_DEV}" ]; then
sudo partx -d "${OPNFV_NBD_DEV}" || true
sudo kpartx -d "${OPNFV_NBD_DEV}" || true
sudo qemu-nbd -d "${OPNFV_NBD_DEV}" || true
fi
}
|