From 51fb22efd1cfa6e415cf46e311a5a660ace3c3e9 Mon Sep 17 00:00:00 2001 From: Jesse Andrews Date: Wed, 19 Oct 2011 09:24:17 -0700 Subject: [PATCH] update comments and added check for double mount bug --- stack.sh | 44 ++++++++++++++++++++++++++++---------------- 1 file changed, 28 insertions(+), 16 deletions(-) diff --git a/stack.sh b/stack.sh index f603ce1..d368293 100755 --- a/stack.sh +++ b/stack.sh @@ -30,18 +30,19 @@ if ! grep -q natty /etc/lsb-release; then fi fi +# Keep track of the current devstack directory. +TOP_DIR=$(cd $(dirname "$0") && pwd) + # stack.sh keeps the list of **apt** and **pip** dependencies in external # files, along with config templates and other useful files. You can find these # in the ``files`` directory (next to this script). We will reference this # directory using the ``FILES`` variable in this script. -FILES=`pwd`/files +FILES=$TOP_DIR/files if [ ! -d $FILES ]; then echo "ERROR: missing devstack/files - did you grab more than just stack.sh?" exit 1 fi -# Keep track of the current devstack directory. -TOP_DIR=$(cd $(dirname "$0") && pwd) # OpenStack is designed to be run as a regular user (Dashboard will fail to run # as root, since apache refused to startup serve content from root user). If @@ -292,6 +293,8 @@ sudo PIP_DOWNLOAD_CACHE=/var/cache/pip pip install `cat $FILES/pips/*` function git_clone { # if there is an existing checkout, move it out of the way if [[ "$RECLONE" == "yes" ]]; then + # FIXME(ja): if we were smarter we could speed up RECLONE by + # using the old git repo as the basis of our new clone... if [ -d $2 ]; then mv $2 /tmp/stack.`date +%s` fi @@ -408,6 +411,8 @@ if [[ "$ENABLED_SERVICES" =~ "dash" ]]; then # ``local_settings.py`` is used to override dashboard default settings. cp $FILES/dash_settings.py $DASH_DIR/openstack-dashboard/local/local_settings.py + # Initialize the dashboard database (it stores sessions and notices shown to + # users). The user system is external (keystone). cd $DASH_DIR/openstack-dashboard dashboard/manage.py syncdb @@ -435,7 +440,8 @@ if [[ "$ENABLED_SERVICES" =~ "g-reg" ]]; then # (re)create glance database mysql -u$MYSQL_USER -p$MYSQL_PASSWORD -e 'DROP DATABASE IF EXISTS glance;' mysql -u$MYSQL_USER -p$MYSQL_PASSWORD -e 'CREATE DATABASE glance;' - # Copy over our glance-registry.conf + + # Copy over our glance configurations and update them GLANCE_CONF=$GLANCE_DIR/etc/glance-registry.conf cp $FILES/glance-registry.conf $GLANCE_CONF sudo sed -e "s,%SQL_CONN%,$BASE_SQL_CONN/glance,g" -i $GLANCE_CONF @@ -454,7 +460,7 @@ fi # We are going to use the sample http middleware configuration from the keystone # project to launch nova. This paste config adds the configuration required # for nova to validate keystone tokens - except we need to switch the config -# to use our admin token instead (instead of the token from their sample data). +# to use our service token instead (instead of the invalid token 999888777666). sudo sed -e "s,999888777666,$SERVICE_TOKEN,g" -i $KEYSTONE_DIR/examples/paste/nova-api-paste.ini if [[ "$ENABLED_SERVICES" =~ "n-cpu" ]]; then @@ -465,9 +471,9 @@ if [[ "$ENABLED_SERVICES" =~ "n-cpu" ]]; then # attempt to load modules: network block device - used to manage qcow images sudo modprobe nbd || true - # Check for kvm (hardware based virtualization). If unable to load kvm, - # set the libvirt type to qemu. Note: many systems come with hardware - # virtualization disabled in BIOS. + # Check for kvm (hardware based virtualization). If unable to initialize + # kvm, we drop back to the slower emulation mode (qemu). Note: many systems + # come with hardware virtualization disabled in BIOS. if [[ "$LIBVIRT_TYPE" == "kvm" ]]; then sudo modprobe kvm || true if [ ! -e /dev/kvm ]; then @@ -481,7 +487,8 @@ if [[ "$ENABLED_SERVICES" =~ "n-cpu" ]]; then # to simulate multiple systems. if [[ "$LIBVIRT_TYPE" == "lxc" ]]; then sudo apt-get install lxc -y - # lxc requires cgroups to be configured on /cgroup + # lxc uses cgroups (a kernel interface via virtual filesystem) configured + # and mounted to ``/cgroup`` sudo mkdir -p /cgroup if ! grep -q cgroup /etc/fstab; then echo none /cgroup cgroup cpuacct,memory,devices,cpu,freezer,blkio 0 0 | sudo tee -a /etc/fstab @@ -491,9 +498,12 @@ if [[ "$ENABLED_SERVICES" =~ "n-cpu" ]]; then fi fi - # User needs to be member of libvirtd group for nova-compute to use libvirt. + # The user that nova runs as needs to be member of libvirtd group otherwise + # nova-compute will be unable to use libvirt. sudo usermod -a -G libvirtd `whoami` - # if kvm wasn't running before we need to restart libvirt to enable it + # libvirt detects various settings on startup, as we potentially changed + # the system configuration (modules, filesystems), we need to restart + # libvirt to detect those changes. sudo /etc/init.d/libvirt-bin restart @@ -503,12 +513,14 @@ if [[ "$ENABLED_SERVICES" =~ "n-cpu" ]]; then # Nova stores each instance in its own directory. mkdir -p $NOVA_DIR/instances - # if there is a partition labeled nova-instances use it (ext filesystems - # can be labeled via e2label) - ## FIXME: if already mounted this blows up... + # You can specify a different disk to be mounted and used for backing the + # virtual machines. If there is a partition labeled nova-instances we + # mount it (ext filesystems can be labeled via e2label). if [ -L /dev/disk/by-label/nova-instances ]; then - sudo mount -L nova-instances $NOVA_DIR/instances - sudo chown -R `whoami` $NOVA_DIR/instances + if ! mount -n | grep -q nova-instances; then + sudo mount -L nova-instances $NOVA_DIR/instances + sudo chown -R `whoami` $NOVA_DIR/instances + fi fi # Clean out the instances directory.