InvenTree/contrib/packager.io/postinstall.sh

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

52 lines
1.6 KiB
Bash
Raw Normal View History

Add package-based install method (#3743) * Add runtime * add procfile * add settings for packaging * use heroku buildpack * maybe a hash works? * try brnach instead of tag or hash * remove reference alltogether * maybe this syntax? * switch back to normal buildpack * and heroku again * try official sample * maybe a tar? * just git clone? * lower version * lower version * switch of settings * try this again * maybe install python before? * no sudo? * just build for ubuntu? * set runtime lower and use custom buildpack * remove initial command * add release command * do not collect static while releaseing * switch to custom backend * Update .pkgr.yml * new backend * upgrade runtime version * consolidate env * update python version * add debian as target * fix os name * extend requirements * remove release command * remove ssl dep * add worker command * add scripts * debug listing * make scripts executable * use python3 * add python install * add python as dep * debug step * switch to app home * set paths right * check local command * remove path * update path * add sample install script * more debugging * dbg msg * more enviroment stuff * more dbugging * more defined install script * reduce function * add debug script * and more * remove before * reduce lines * cleanup deps * maybe this way? * rm " * rename env * fix files * use dot annotation * ls * add msgs * move activation * ref? * use source * use point annotation * remove activate * create env and add it to the enviroment * use full paths * move python init * use modern python * use version numbers for ubuntu * add default env vars ( for now) * use export * use shorthand command * add more variables * use apphome reference in paths * add wasyprint dependencies * also use wheels * add gettext for comiling messages * add nginx setting * remove postgres from required stack * refactor fiel configs * make sure parent dirs exsist * add wheels * move python setup again * update packe name * update gunicron command * move to config * try cleaner procfile * use python 3 wheels * update service def * Add enviroment settings to packaging * restart nginx * remove slack * change nginx config name * fix styles * set permissions on update * Add single line install to README * move nginx names back * use production sample as nginx config * fix sed statement * cleaner nginx setup * keep configs in config file instead of packaging index * fix sed connamds to remove spaces * reintroduce spaces * Also disable debug mode * fix file permissions * remove unneeded commands * readd missing spaces * add admin user generation * Show public ipadress in final message * small cleanups * Add log to setup * remove logging * remove logging parts * chown all dirs * run update command as right user already * set path before running command * fix permission setup * switch right assignment * run commands with sudo * pass in envs * preserve only selected enviroment * show enviroment variables * silence curl * fix envs * clean up final message * move nginx file to variable * make admin creation conditional * a bit more overview for setup variables * test if nginx is available * unlink instead of delet * clean up functions * well that was dumb * ok double dumb - now i tested it * use config for admin generation * fix logic * show value - not call it * switch functions to sh * fix syntax * add admin data to exposed envs * remove keeping of admin pwd * switch to true/false * make nginx file location configureable * use gunicorn instead * fix syntax * set base dir * move env to seperate function * Also use/write database settings * use conf values if available * format sh * update passed envs * fix dep * enable installing extra packages via env (ie database backends) * ann more logging steps * add more debuggin steps * get local envs * add more debug and reorder * more code for testing * fix call * Add debug msg * turn install quiet * fix syntax * reduce debugging * add log step * add datestamp * add marker to inv execution * fix syntax * add codeowner
2022-10-15 23:14:18 +00:00
#!/bin/bash
#
# packager.io postinstall script
#
exec > >(tee ${APP_HOME}/log/setup_$(date +"%F_%H_%M_%S").log) 2>&1
PATH=${APP_HOME}/env/bin:${APP_HOME}/:/sbin:/bin:/usr/sbin:/usr/bin:
# import functions
. ${APP_HOME}/contrib/packager.io/functions.sh
# Envs that should be passed to setup commands
export SETUP_ENVS=PATH,APP_HOME,INVENTREE_MEDIA_ROOT,INVENTREE_STATIC_ROOT,INVENTREE_BACKUP_DIR,INVENTREE_PLUGINS_ENABLED,INVENTREE_PLUGIN_FILE,INVENTREE_CONFIG_FILE,INVENTREE_SECRET_KEY_FILE,INVENTREE_DB_ENGINE,INVENTREE_DB_NAME,INVENTREE_DB_USER,INVENTREE_DB_PASSWORD,INVENTREE_DB_HOST,INVENTREE_DB_PORT,INVENTREE_ADMIN_USER,INVENTREE_ADMIN_EMAIL,INVENTREE_ADMIN_PASSWORD,SETUP_NGINX_FILE,SETUP_ADMIN_PASSWORD_FILE,SETUP_NO_CALLS,SETUP_DEBUG,SETUP_EXTRA_PIP,SETUP_PYTHON
Add package-based install method (#3743) * Add runtime * add procfile * add settings for packaging * use heroku buildpack * maybe a hash works? * try brnach instead of tag or hash * remove reference alltogether * maybe this syntax? * switch back to normal buildpack * and heroku again * try official sample * maybe a tar? * just git clone? * lower version * lower version * switch of settings * try this again * maybe install python before? * no sudo? * just build for ubuntu? * set runtime lower and use custom buildpack * remove initial command * add release command * do not collect static while releaseing * switch to custom backend * Update .pkgr.yml * new backend * upgrade runtime version * consolidate env * update python version * add debian as target * fix os name * extend requirements * remove release command * remove ssl dep * add worker command * add scripts * debug listing * make scripts executable * use python3 * add python install * add python as dep * debug step * switch to app home * set paths right * check local command * remove path * update path * add sample install script * more debugging * dbg msg * more enviroment stuff * more dbugging * more defined install script * reduce function * add debug script * and more * remove before * reduce lines * cleanup deps * maybe this way? * rm " * rename env * fix files * use dot annotation * ls * add msgs * move activation * ref? * use source * use point annotation * remove activate * create env and add it to the enviroment * use full paths * move python init * use modern python * use version numbers for ubuntu * add default env vars ( for now) * use export * use shorthand command * add more variables * use apphome reference in paths * add wasyprint dependencies * also use wheels * add gettext for comiling messages * add nginx setting * remove postgres from required stack * refactor fiel configs * make sure parent dirs exsist * add wheels * move python setup again * update packe name * update gunicron command * move to config * try cleaner procfile * use python 3 wheels * update service def * Add enviroment settings to packaging * restart nginx * remove slack * change nginx config name * fix styles * set permissions on update * Add single line install to README * move nginx names back * use production sample as nginx config * fix sed statement * cleaner nginx setup * keep configs in config file instead of packaging index * fix sed connamds to remove spaces * reintroduce spaces * Also disable debug mode * fix file permissions * remove unneeded commands * readd missing spaces * add admin user generation * Show public ipadress in final message * small cleanups * Add log to setup * remove logging * remove logging parts * chown all dirs * run update command as right user already * set path before running command * fix permission setup * switch right assignment * run commands with sudo * pass in envs * preserve only selected enviroment * show enviroment variables * silence curl * fix envs * clean up final message * move nginx file to variable * make admin creation conditional * a bit more overview for setup variables * test if nginx is available * unlink instead of delet * clean up functions * well that was dumb * ok double dumb - now i tested it * use config for admin generation * fix logic * show value - not call it * switch functions to sh * fix syntax * add admin data to exposed envs * remove keeping of admin pwd * switch to true/false * make nginx file location configureable * use gunicorn instead * fix syntax * set base dir * move env to seperate function * Also use/write database settings * use conf values if available * format sh * update passed envs * fix dep * enable installing extra packages via env (ie database backends) * ann more logging steps * add more debuggin steps * get local envs * add more debug and reorder * more code for testing * fix call * Add debug msg * turn install quiet * fix syntax * reduce debugging * add log step * add datestamp * add marker to inv execution * fix syntax * add codeowner
2022-10-15 23:14:18 +00:00
# Get the envs
detect_local_env
# default config
export CONF_DIR=/etc/inventree
export DATA_DIR=${APP_HOME}/data
# Setup variables
export SETUP_NGINX_FILE=${SETUP_NGINX_FILE:-/etc/nginx/sites-enabled/inventree.conf}
export SETUP_ADMIN_PASSWORD_FILE=${CONF_DIR}/admin_password.txt
export SETUP_NO_CALLS=${SETUP_NO_CALLS:-false}
export SETUP_PYTHON=${SETUP_PYTHON:-python3.9}
Add package-based install method (#3743) * Add runtime * add procfile * add settings for packaging * use heroku buildpack * maybe a hash works? * try brnach instead of tag or hash * remove reference alltogether * maybe this syntax? * switch back to normal buildpack * and heroku again * try official sample * maybe a tar? * just git clone? * lower version * lower version * switch of settings * try this again * maybe install python before? * no sudo? * just build for ubuntu? * set runtime lower and use custom buildpack * remove initial command * add release command * do not collect static while releaseing * switch to custom backend * Update .pkgr.yml * new backend * upgrade runtime version * consolidate env * update python version * add debian as target * fix os name * extend requirements * remove release command * remove ssl dep * add worker command * add scripts * debug listing * make scripts executable * use python3 * add python install * add python as dep * debug step * switch to app home * set paths right * check local command * remove path * update path * add sample install script * more debugging * dbg msg * more enviroment stuff * more dbugging * more defined install script * reduce function * add debug script * and more * remove before * reduce lines * cleanup deps * maybe this way? * rm " * rename env * fix files * use dot annotation * ls * add msgs * move activation * ref? * use source * use point annotation * remove activate * create env and add it to the enviroment * use full paths * move python init * use modern python * use version numbers for ubuntu * add default env vars ( for now) * use export * use shorthand command * add more variables * use apphome reference in paths * add wasyprint dependencies * also use wheels * add gettext for comiling messages * add nginx setting * remove postgres from required stack * refactor fiel configs * make sure parent dirs exsist * add wheels * move python setup again * update packe name * update gunicron command * move to config * try cleaner procfile * use python 3 wheels * update service def * Add enviroment settings to packaging * restart nginx * remove slack * change nginx config name * fix styles * set permissions on update * Add single line install to README * move nginx names back * use production sample as nginx config * fix sed statement * cleaner nginx setup * keep configs in config file instead of packaging index * fix sed connamds to remove spaces * reintroduce spaces * Also disable debug mode * fix file permissions * remove unneeded commands * readd missing spaces * add admin user generation * Show public ipadress in final message * small cleanups * Add log to setup * remove logging * remove logging parts * chown all dirs * run update command as right user already * set path before running command * fix permission setup * switch right assignment * run commands with sudo * pass in envs * preserve only selected enviroment * show enviroment variables * silence curl * fix envs * clean up final message * move nginx file to variable * make admin creation conditional * a bit more overview for setup variables * test if nginx is available * unlink instead of delet * clean up functions * well that was dumb * ok double dumb - now i tested it * use config for admin generation * fix logic * show value - not call it * switch functions to sh * fix syntax * add admin data to exposed envs * remove keeping of admin pwd * switch to true/false * make nginx file location configureable * use gunicorn instead * fix syntax * set base dir * move env to seperate function * Also use/write database settings * use conf values if available * format sh * update passed envs * fix dep * enable installing extra packages via env (ie database backends) * ann more logging steps * add more debuggin steps * get local envs * add more debug and reorder * more code for testing * fix call * Add debug msg * turn install quiet * fix syntax * reduce debugging * add log step * add datestamp * add marker to inv execution * fix syntax * add codeowner
2022-10-15 23:14:18 +00:00
# SETUP_DEBUG can be set to get debug info
# SETUP_EXTRA_PIP can be set to install extra pip packages
# SETUP_PYTHON can be set to use a different python version
Add package-based install method (#3743) * Add runtime * add procfile * add settings for packaging * use heroku buildpack * maybe a hash works? * try brnach instead of tag or hash * remove reference alltogether * maybe this syntax? * switch back to normal buildpack * and heroku again * try official sample * maybe a tar? * just git clone? * lower version * lower version * switch of settings * try this again * maybe install python before? * no sudo? * just build for ubuntu? * set runtime lower and use custom buildpack * remove initial command * add release command * do not collect static while releaseing * switch to custom backend * Update .pkgr.yml * new backend * upgrade runtime version * consolidate env * update python version * add debian as target * fix os name * extend requirements * remove release command * remove ssl dep * add worker command * add scripts * debug listing * make scripts executable * use python3 * add python install * add python as dep * debug step * switch to app home * set paths right * check local command * remove path * update path * add sample install script * more debugging * dbg msg * more enviroment stuff * more dbugging * more defined install script * reduce function * add debug script * and more * remove before * reduce lines * cleanup deps * maybe this way? * rm " * rename env * fix files * use dot annotation * ls * add msgs * move activation * ref? * use source * use point annotation * remove activate * create env and add it to the enviroment * use full paths * move python init * use modern python * use version numbers for ubuntu * add default env vars ( for now) * use export * use shorthand command * add more variables * use apphome reference in paths * add wasyprint dependencies * also use wheels * add gettext for comiling messages * add nginx setting * remove postgres from required stack * refactor fiel configs * make sure parent dirs exsist * add wheels * move python setup again * update packe name * update gunicron command * move to config * try cleaner procfile * use python 3 wheels * update service def * Add enviroment settings to packaging * restart nginx * remove slack * change nginx config name * fix styles * set permissions on update * Add single line install to README * move nginx names back * use production sample as nginx config * fix sed statement * cleaner nginx setup * keep configs in config file instead of packaging index * fix sed connamds to remove spaces * reintroduce spaces * Also disable debug mode * fix file permissions * remove unneeded commands * readd missing spaces * add admin user generation * Show public ipadress in final message * small cleanups * Add log to setup * remove logging * remove logging parts * chown all dirs * run update command as right user already * set path before running command * fix permission setup * switch right assignment * run commands with sudo * pass in envs * preserve only selected enviroment * show enviroment variables * silence curl * fix envs * clean up final message * move nginx file to variable * make admin creation conditional * a bit more overview for setup variables * test if nginx is available * unlink instead of delet * clean up functions * well that was dumb * ok double dumb - now i tested it * use config for admin generation * fix logic * show value - not call it * switch functions to sh * fix syntax * add admin data to exposed envs * remove keeping of admin pwd * switch to true/false * make nginx file location configureable * use gunicorn instead * fix syntax * set base dir * move env to seperate function * Also use/write database settings * use conf values if available * format sh * update passed envs * fix dep * enable installing extra packages via env (ie database backends) * ann more logging steps * add more debuggin steps * get local envs * add more debug and reorder * more code for testing * fix call * Add debug msg * turn install quiet * fix syntax * reduce debugging * add log step * add datestamp * add marker to inv execution * fix syntax * add codeowner
2022-10-15 23:14:18 +00:00
# get base info
detect_envs
detect_docker
detect_initcmd
detect_ip
# create processes
create_initscripts
create_admin
# run updates
stop_inventree
update_or_install
# Write config file
if [ "${SETUP_CONF_LOADED}" = "true" ]; then
set_env
fi
start_inventree
# show info
final_message