Improve logging for deploy start shell scripts

This commit improves logging during deploy start by:
1. Creating a common module to be sourced by shell scripts to
   load general-use functions, thus reducing code duplication
   between the scripts
2. Replacing plain "echo" commands on the scripts by logging
   functions present on the common module
3. Adding timestamps to the log messages
4. Centralizing all scripts logs into software.log, favouring
   the troubleshooting, now that log lines contain timestamps
   and the process/script that generated them

This commit also deletes the ostree_mounts.yaml file since
it would be used by apt-ostree integration, which was dropped.

Test Plan
PASS: run deploy start successfully and verify that deploy start
      log messages are logged with the expected format

Story: 2010676
Task: 49607

Change-Id: I0bdebde8147faa5b29a642e35bfaf26e9862ed0a
Signed-off-by: Heitor Matsui <heitorvieira.matsui@windriver.com>
This commit is contained in:
Heitor Matsui 2024-02-22 11:13:59 -03:00
parent e96e621b6d
commit af0a86e357
7 changed files with 123 additions and 60 deletions

View File

@ -74,8 +74,6 @@ override_dh_install:
${ROOT}/etc/software/${METADATA_FILE}
install -m 755 scripts/chroot_mounts.sh \
${ROOT}/usr/sbin/software-deploy/chroot_mounts.sh
install -m 444 scripts/ostree_mounts.yaml \
${ROOT}/etc/software/ostree_mounts.yaml
install -m 755 scripts/create_postgresql_database.sh \
${ROOT}/usr/sbin/software-deploy/create_postgresql_database.sh
install -m 755 scripts/usm_load_import \
@ -84,4 +82,6 @@ override_dh_install:
${ROOT}/usr/sbin/software-deploy/sync-controllers-feed
install -m 755 scripts/deploy-cleanup \
${ROOT}/usr/sbin/software-deploy/deploy-cleanup
install -m 755 scripts/shell-utils \
${ROOT}/usr/sbin/software-deploy/shell-utils
dh_install

View File

@ -5,6 +5,15 @@
# SPDX-License-Identifier: Apache-2.0
#
script_dir=$(dirname $0)
shell_utils=${script_dir}/shell-utils
if [ -f $shell_utils ]; then
source $shell_utils
else
echo "ERROR: ${shell_utils} module not found."
exit 1
fi
DEV_PATH=/dev
PLATFORM_PATH=/opt/platform
RABBIT_PATH=/var/lib/rabbitmq
@ -34,8 +43,8 @@ handle_error() {
local exit_code="$1"
local error_message="$2"
echo "Error: $error_message" >&2
echo "Please check the error details and take appropriate action for recovery." >&2
log "$error_message" >&2
log "Please check the error details and take appropriate action for recovery." >&2
# attempt to unmount if there were successful mounts before the error
umount_all
@ -50,7 +59,7 @@ mount_all() {
src=${src_dst[0]}
dst=${src_dst[1]}
echo "mount --bind ${src} ${dst}"
info "Bind mounting ${src} -> ${dst}"
sudo mkdir -p ${dst}
sudo mount --bind ${src} ${dst} || handle_error 1 "Failed to bind mount ${src} to ${dst}"
done
@ -65,12 +74,12 @@ umount_all() {
src=${src_dst[0]}
dst=${src_dst[1]}
echo "sudo umount $dst"
info "Unmounting $dst"
umount_output=$(sudo umount $dst 2>&1)
if [ $? -ne 0 ]; then
# ignore messages that are not harmful
if [[ ! $umount_output =~ ("not mounted"|"no mount point specified") ]]; then
echo $umount_output
error "$umount_output"
rc=1
fi
fi
@ -94,16 +103,16 @@ check_all() {
fi
done
if [[ ${#mounted[@]} -gt 0 ]]; then
echo "Mounted mount points:"
info "Mounted mount points:"
for mnt in ${mounted[@]}; do
echo $mnt
info $mnt
done
fi
return $rc
}
if [ -z "$1" ]; then
echo "Error: OSTree deployment branch parameter is missing."
error "OSTree deployment branch parameter is missing."
exit 1
fi

View File

@ -1,13 +1,22 @@
#!/bin/bash
#
# Copyright (c) 2023 Wind River Systems, Inc.
# Copyright (c) 2023-2024 Wind River Systems, Inc.
#
# SPDX-License-Identifier: Apache-2.0
#
#
# This script is used create a 2nd instance of postgres on a DX upgrade.
# It needs the port number as parameter and it should be different from the default.
#
script_dir=$(dirname $0)
shell_utils=${script_dir}/shell-utils
if [ -f $shell_utils ]; then
source $shell_utils
else
echo "ERROR: ${shell_utils} module not found."
exit 1
fi
DEFAULT_POSTGRESQL_PORT=5432
POSTGRESQL_PATH=/var/lib/postgresql
@ -16,7 +25,7 @@ POSTGRESQL_RUNTIME=/var/run/postgresql
INFO_FILE=/etc/build.info
if [ -z "$1" ]; then
echo "Error: Port parameter is missing."
error "'Port' parameter is missing."
exit 1
fi
@ -24,7 +33,7 @@ PORT="$1"
# Prevent issues with the default postgres port
if [ "$PORT" -eq "$DEFAULT_POSTGRESQL_PORT" ]; then
echo "Error: Port number should be different from the default."
error "Port number should be different from the default."
exit 1
fi
@ -32,8 +41,8 @@ cleanup_and_exit() {
local exit_code="$1"
local error_message="$2"
echo "Error: $error_message" >&2
echo "Please check the error details and take appropriate action for recovery." >&2
error "$error_message" >&2
error "Please check the error details and take appropriate action for recovery." >&2
exit "$exit_code"
}

View File

@ -15,10 +15,14 @@
# automatic cleanup process fails.
#
log() {
type=$1
echo "[$(date --iso-8601=seconds)] $type: $2"
}
script_dir=$(dirname $0)
shell_utils=${script_dir}/shell-utils
if [ -f $shell_utils ]; then
source $shell_utils
else
echo "ERROR: ${shell_utils} module not found."
exit 1
fi
stop_database() {
local rootdir=$1
@ -27,7 +31,7 @@ stop_database() {
# attempt to stop temporary database if still up
tmp_db_dir=${rootdir}/var/lib/postgresql/${to_ver}
log "INFO" "Attempting to stop the temporary database in ${tmp_db_dir}..."
info "Attempting to stop the temporary database in ${tmp_db_dir}..."
if [ -d $tmp_db_dir ]; then
lsof $tmp_db_dir
if [ $? -eq 0 ]; then
@ -35,15 +39,15 @@ stop_database() {
sudo -u postgres ${tmp_db_bin_dir}/pg_ctl -D ${tmp_db_dir} stop
if [ $? -ne 0 ]; then
rc=1
log "ERROR" "Error stopping database."
error "Error stopping database."
else
log "INFO" "Success stopping database."
info "Success stopping database."
fi
else
log "INFO" "Database is not running."
info "Database is not running."
fi
else
log "WARN" "No database found in the specified directory."
warning "No database found in the specified directory."
fi
return $rc
}
@ -52,20 +56,20 @@ unmount_filesystems() {
local rootdir=$1
local rc=0
log "INFO" "Attempting to unmount filesystems under ${rootdir}..."
info "Attempting to unmount filesystems under ${rootdir}..."
k8s_umount_output=$(sudo umount ${rootdir}/usr/local/kubernetes/current 2>&1)
if [ $? -ne 0 ]; then
if [[ ! $k8s_umount_output =~ "not mounted" ]]; then
rc=1
log "ERROR" $k8s_umount_output
error $k8s_umount_output
fi
fi
sudo ${rootdir}/usr/sbin/software-deploy/chroot_mounts.sh ${rootdir} -u
if [ $? -ne 0 ]; then
rc=1
log "ERROR" "Error unmounting filesystems."
error "Error unmounting filesystems."
else
log "INFO" "Success unmounting filesystems."
info "Success unmounting filesystems."
fi
return $rc
}
@ -75,14 +79,14 @@ remove_temp_directories() {
local rootdir=$2
local rc=0
log "INFO" "Attempting to remove temporary deployment directories [${repo}, ${rootdir}]..."
info "Attempting to remove temporary deployment directories [${repo}, ${rootdir}]..."
sudo ${rootdir}/usr/sbin/software-deploy/chroot_mounts.sh ${rootdir} -c
if [ $? -ne 0 ]; then
rc=1
log "ERROR" "Some mount points are still mounted, cannot proceed with the cleanup."
error "Some mount points are still mounted, cannot proceed with the cleanup."
else
rm -rf $repo $rootdir
log "INFO" "Temporary deployment directories removed successfully."
info "Temporary deployment directories removed successfully."
fi
return $rc
}
@ -90,7 +94,7 @@ remove_temp_directories() {
# script usage
if [ $# -ne 3 ]; then
echo
echo "usage: deploy-start-cleanup <tmp_ostree_repo_dir> <tmp_root_dir> <action>"
echo "usage: deploy-cleanup <tmp_ostree_repo_dir> <tmp_root_dir> <action>"
echo -e "\nParameters"
echo "=========="
echo "tmp_ostree_repo_dir: temporary ostree repo directory (example: /sysroot/upgrade/ostree_repo)"
@ -111,14 +115,14 @@ action=$3
# basic checks
for dir in $repo $rootdir; do
if [ ! -d $dir ]; then
log "ERROR" "Specified directory $dir does not exist, cannot proceed with cleanup."
error "Specified directory $dir does not exist, cannot proceed with cleanup."
exit 1
fi
done
target_build_info=${rootdir}/usr/etc/build.info
if [ ! -f $target_build_info ]; then
log "ERROR" "Cannot get target release build-info information, cannot proceed with cleanup."
error "Cannot get target release build-info information, cannot proceed with cleanup."
exit 1
fi
@ -126,7 +130,7 @@ fi
to_ver=$(cat $target_build_info | grep SW_VERSION | cut -d'"' -f2)
# main
log "INFO" "Starting cleanup for staging directories [${repo}, ${rootdir}]..."
info "Starting cleanup for staging directories [${repo}, ${rootdir}]..."
case $action in
"db")
stop_database $rootdir $to_ver
@ -143,13 +147,13 @@ case $action in
remove_temp_directories $repo $rootdir
;;
*)
log "ERROR" "Invalid action specified: ${action}"
error "Invalid action specified: ${action}"
;;
esac
rc=$?
if [ $rc -ne 0 ]; then
log "ERROR" "Error cleaning up [${repo}, ${rootdir}], please check the logs, take manual actions and retry the script."
error "Error cleaning up [${repo}, ${rootdir}], please check the logs, take manual actions and retry the script."
fi
log "INFO" "Cleanup script ended."
info "Cleanup script ended."
exit $rc

View File

@ -1,10 +0,0 @@
---
- "{{ OSTREE_DEPLOYMENT_BRANCH }}/usr/etc:{{ OSTREE_DEPLOYMENT_BRANCH }}/etc"
- "/dev:{{ OSTREE_DEPLOYMENT_BRANCH }}/dev"
- "/proc:{{ OSTREE_DEPLOYMENT_BRANCH }}/proc"
- "/opt/platform:{{ OSTREE_DEPLOYMENT_BRANCH }}/opt/platform"
- "/var/lib/rabbitmq:{{ OSTREE_DEPLOYMENT_BRANCH }}/var/lib/rabbitmq"
- "/var/log:{{ OSTREE_DEPLOYMENT_BRANCH }}/var/log"
- >
"/var/lib/postgresql/upgrade:
{{ OSTREE_DEPLOYMENT_BRANCH }}/var/lib/postgresql/upgrade"

View File

@ -0,0 +1,29 @@
# !/bin/bash
#
# Copyright (c) 2024 Wind River Systems, Inc.
#
# SPDX-License-Identifier: Apache-2.0
#
# File containing common shell functions that
# can be sourced and used by other shell scripts.
#
log()
{
script_name=$(basename $0)
log_type=$1
shift
echo "$(date -Iseconds | cut -d'+' -f1): ${script_name}[${$}]: ${log_type}: ${@}"
}
info() {
log "INFO" $@
}
warning() {
log "WARNING" $@
}
error() {
log "ERROR" $@ >&2
}

View File

@ -13,10 +13,17 @@
# 5. perform data migration
#
# TODO: centralize USM upgrade scripts output into one single log
exec > /var/log/deploy_start.log 2>&1
exec 2>&1 >> /var/log/software.log
script_dir=$(dirname $0)
shell_utils=${script_dir}/shell-utils
if [ -f $shell_utils ]; then
source $shell_utils
else
echo "ERROR: ${shell_utils} module not found."
exit 1
fi
exec_path=$(dirname $0)
usage()
{
echo "usage: $0 from_ver to_ver k8s_ver postgresql_port feed [commit_id|latest_commit]"
@ -63,10 +70,10 @@ handle_error() {
local error_message="$2"
local state="start-failed"
echo "Error: ${error_message}" >&2
echo "Please check the error details and take appropriate action for recovery." >&2
error "${error_message}" >&2
error "Please check the error details and take appropriate action for recovery." >&2
echo "Update deploy state ${state}." >&2
error "Update deploy state ${state}." >&2
deploy_update_state ${state}
# cleanup before exiting
@ -77,13 +84,14 @@ handle_error() {
for dir in $rootdir $repo; do
if [ -e ${dir} ]; then
echo "${dir} already exists. Please ensure to clean up environment to continue." >&2
error "${dir} already exists. Please ensure to clean up environment to continue." >&2
exit 1
fi
done
sudo mkdir ${repo} -p
info "Initializing ostree repo..."
sudo ostree --repo=${repo} init --mode=archive || handle_error $? "Failed to init repo"
sudo ostree --repo=${repo} remote add ${instbr} ${feed_url} --no-gpg-verify || handle_error $? "Failed to remote add repo"
sudo ostree --repo=${repo} pull --depth=-1 --mirror ${instbr}:${instbr} || handle_error $? "Failed to pull repo"
@ -92,23 +100,27 @@ sudo ostree --repo=${repo} pull --depth=-1 --mirror ${instbr}:${instbr} || handl
if [ -z ${commit_id} ]; then
# get commit id, only latest for now
commit_id=$(ostree rev-parse --repo=${repo} ${instbr})
echo "latest commit id ${commit_id}"
info "Latest commit id ${commit_id}"
if [ -z "${commit_id}" ]; then
handle_error 1 "Failed to retrieve commit id"
fi
fi
echo "checkout commit id ${commit_id}"
info "Checkout commit id ${commit_id}"
info "Checking out ostree repo..."
sudo ostree --repo=${repo} checkout ${commit_id} ${rootdir} || handle_error $? "Failed to checkout ${commit_id}"
# create proper mounts on deploy file system
info "Creating mount points..."
sudo ${rootdir}/usr/sbin/software-deploy/chroot_mounts.sh ${rootdir} || handle_error $? "Failed to mount required mount points"
sudo mount --bind ${rootdir}/usr/local/kubernetes/${k8s_ver} ${rootdir}/usr/local/kubernetes/current
sudo cp /etc/kubernetes/admin.conf ${rootdir}/etc/kubernetes/
info "Mount points created successfully"
DATA_PREP_SCRIPT="${exec_path}/prep-data-migration"
info "Preparing for data migration..."
DATA_PREP_SCRIPT="${script_dir}/prep-data-migration"
# OS_AUTH_URL, OS_USERNAME, OS_PASSWORD, OS_PROJECT_NAME, OS_USER_DOMAIN_NAME,
# OS_PROJECT_DOMAIN_NAME, OS_REGION_NAME are in env variables.
cmd_line=" --rootdir=${rootdir} --from_release=${from_ver} --to_release=${to_ver}"
@ -118,17 +130,27 @@ cmd_line+=" --project_domain_name=${OS_PROJECT_DOMAIN_NAME} --region_name=${OS_R
prep_cmd="${DATA_PREP_SCRIPT} ${cmd_line}"
${prep_cmd} || handle_error $? "Failed to extract data for migration"
info "Data migration preparations complete."
info "Creating temporary database..."
sudo chroot ${rootdir} /usr/sbin/software-deploy/create_postgresql_database.sh ${port} || handle_error $? "Failed to start 2nd instance of postgresql"
sudo chroot ${rootdir} /usr/bin/software-migrate ${from_ver} ${to_ver} ${port} || handle_error $? "Failed to migrate data"
info "Database creation complete."
info "Starting data migration..."
sudo chroot ${rootdir} /usr/bin/software-migrate ${from_ver} ${to_ver} ${port} || handle_error $? "Failed to migrate data"
info "Data migration completed."
info "Syncing feed between controllers..."
SYNC_CONTROLLERS_SCRIPT="/usr/sbin/software-deploy/sync-controllers-feed"
sync_controllers_cmd="${SYNC_CONTROLLERS_SCRIPT} ${cmd_line} --feed=${feed}"
${sync_controllers_cmd} || handle_error $? "Failed to sync feeds"
info "Feed sync complete."
state="start-done"
deploy_update_state $state
echo "Update deploy state ${state}."
info "Update deploy state ${state}."
# cleanup after successful data migration
info "Starting cleanup..."
deploy_cleanup
info "Cleanup complete."