Down very slow?

[filisko]

Hi guys

For some reason. when I do docker compose down it takes quite a while when for me it should be instant.

This is my dockerfile:

# https://serversideup.net/open-source/docker-php/
FROM serversideup/php:8.4-fpm-nginx AS base
ENV ROOT_PATH=/var/www/html

USER root
RUN install-php-extensions gd sockets intl igbinary msgpack relay

WORKDIR /var/www/html

############################################
# Development Image
############################################
FROM base AS development
ENV ENVIRONMENT=development

ARG USER_ID
ARG GROUP_ID

RUN echo "permisions $USER_ID:$GROUP_ID"

USER root
RUN docker-php-serversideup-set-id www-data $USER_ID:$GROUP_ID
RUN install-php-extensions xdebug

RUN chown -R www-data:www-data /etc/nginx
RUN chown -R www-data:www-data /var/cache/nginx
RUN chown -R www-data:www-data /var/log/nginx

USER www-data

############################################
# CI image
############################################
FROM base AS ci
ENV ENVIRONMENT=testing
ENV PHP_OPCACHE_ENABLE=1

# Sometimes CI images need to run as root
# so we set the ROOT user and configure
# the PHP-FPM pool to run as www-data
USER root

RUN echo "user = www-data" >> /usr/local/etc/php-fpm.d/docker-php-serversideup-pool.conf && \
    echo "group = www-data" >> /usr/local/etc/php-fpm.d/docker-php-serversideup-pool.conf

############################################
# Production Image
############################################
FROM base AS deploy
ENV ENVIRONMENT=production
ENV PHP_OPCACHE_ENABLE=1

COPY --chown=www-data:www-data composer.json composer.lock /var/www/html/

USER www-data

RUN composer install \
--no-dev \
--optimize-autoloader \
--classmap-authoritative \
--no-interaction \
--no-progress

COPY --chown=www-data:www-data . /var/www/html

[jaydrogers]

It's probably related to the stop signal stuff:

docker-php/src/variations/fpm-nginx/Dockerfile

Lines 185 to 186 in ae32ba0

	# Set stop signal to SIGQUIT for a graceful shutdown instead of S6's preferred SIGTERM (https://github.com/just-containers/s6-overlay/issues/586)
	STOPSIGNAL SIGQUIT

Here's the full story if you're interested in reading about it: just-containers/s6-overlay#586

Long story short (for fpm-nginx)

• I use stop_signal: SIGTERM when I want things to go down quickly
• Use the default SIGQUIT if you want things to go down without any downtime (like on a zero-downtime deployment)

Other note

• I like to override the stop signal in my orchestrator (like Docker Compose, Docker Swarm, K8s, etc) instead of in the Dockerfile -- but it's your preference where you want to place it 😃

[filisko]

thank you @jaydrogers !!

Unfortunately, that didn't solve the problem. Regarding the stop_signal: SIGTERM. Is that the same as running kill --signal SIGTERM ?

[jaydrogers]

Can you give me the command you're using to bring up your containers so I can replicate on my end?