当我启动我的 docker 容器时，Cron 没有运行

Question

To preface, I have been referencing these two articles for help:

• Run a cron job with Docker - Julien Boulay
• Running cron jobs inside a Docker container - Chris S.

My goal is to have a cron job automatically start when I start my docker container. Currently, it doesn't automatically start, but I can manually go into my container and run service cron start , which starts the job, and it works correctly.

So the problem is: How do I get my cron job to start automatically when my container starts up?

Dockerfile

FROM microsoft/dotnet:latest
RUN apt-get update && apt-get install -y cron

COPY . /app

WORKDIR /app

ADD crontab /etc/cron.d/crontab
RUN chmod 0600 /etc/cron.d/crontab
RUN crontab -u root /etc/cron.d/crontab
RUN touch /var/log/cron.log

RUN ["dotnet", "restore"]

RUN ["dotnet", "build"]

EXPOSE 5000/tcp

CMD cron && tail -f /var/log/cron.log
CMD service cron start

crontab

* * * * * echo "Hello world" >> /var/log/cron.log 2>&1
# Empty space

---

Though I wasn't able to get cron working in that particular container, I was able to create a standalone docker container specifically for cron, and was successful in getting it to run automatically.

As far as setup for the cron container, I followed the linked article, Run a cron job with Docker - Julien Boulay , and was able to get it working.

Answer 1

What I'm doing is have the CMD call cron directly like this:

CMD /usr/sbin/cron -f

Before that I'm adding the crontab to the container and assigning it as the root crontab with the command:

RUN crontab /root/mycrontab

You don't need to call the crontab command on files that are located in /etc/cron.d , but you do need those files to have the correct syntax. Using your example, instead of this:

* * * * * echo "Hello world" >> /var/log/cron.log 2>&1

You should have this:

* * * * * root echo "Hello world" >> /var/log/cron.log 2>&1

On your crontab file. This only applies to crontab files located within /etc/cron.d , otherwise your crontab file syntax is correct and you use the crontab command to load it.

Starting from your example, I think you should modify your files like this:

Dockerfile

FROM microsoft/dotnet:latest
RUN apt-get update && apt-get install -y cron

COPY . /app

WORKDIR /app

ADD crontab /etc/cron.d/crontab
RUN chmod 0600 /etc/cron.d/crontab
RUN touch /var/log/cron.log

RUN ["dotnet", "restore"]

RUN ["dotnet", "build"]

EXPOSE 5000/tcp

CMD /usr/sbin/cron -f

crontab

* * * * * root echo "Hello world" >> /var/log/cron.log 2>&1

Another alternative would be:

Dockerfile

FROM microsoft/dotnet:latest
RUN apt-get update && apt-get install -y cron

COPY . /app

WORKDIR /app

ADD crontab /root/
RUN crontab /root/crontab
RUN touch /var/log/cron.log

RUN ["dotnet", "restore"]

RUN ["dotnet", "build"]

EXPOSE 5000/tcp

CMD /usr/sbin/cron -f

crontab

* * * * * echo "Hello world" >> /var/log/cron.log 2>&1

Answer 2

We had a problem with php-fpm and docker where our cronjob tasks were not be executed. There were two problems we solved:

• We tried to copy a crontab file into the docker container by using COPY config/custom-cron /etc/cron.d/custom-cron . The problem is, that our line endings were in windows format. This did break our crontab file, because this line endings are not converted while copy that file into the container.
• The second problem was, that we tried to start the cron via CMD ["cron", "-f"] which did block the main php-fpm process. This results in a 502 Bad gateway error when calling our web application.

---

Finaly we made it work by editing the crontab file manually while building the docker image instead of copy-pasting and using supervisord to have multiple tasks running inside docker. This should work on all supported operating systems.

dockerfile

FROM php:7.1.16-fpm

RUN apt-get update && apt-get install -y cron supervisor

# Configure cron
RUN crontab -l | { cat; echo "* * * * * echo 'Hello world' >> /var/log/cron-test.log 2>&1"; } | crontab -

# Configure supervisor
COPY config/supervisord.conf /etc/supervisor/supervisord.conf

supervisord.conf

[supervisord]
logfile = /dev/null
loglevel = info
pidfile = /var/run/supervisord.pid
nodaemon = true

[program:php-fpm]
command = php-fpm
autostart = true
autorestart = true
stdout_logfile = /dev/stdout
stdout_logfile_maxbytes = 0
stderr_logfile = /dev/stderr
stderr_logfile_maxbytes = 0

[program:cron]
command = cron -f
autostart = true
autorestart = true
stdout_logfile = /dev/stdout
stdout_logfile_maxbytes = 0
stderr_logfile = /dev/stderr
stderr_logfile_maxbytes = 0

Answer 3

There is a bug in Debian based distributions which will cause cronjobs to fail because docker uses layered filesystem and cron doesn't start and says NUMBER OF HARD LINKS > 1 (/etc/crontab) .

The fix is simple, add touch /etc/crontab /etc/cron.*/* to the entrypoint of your container.

I have made a blog post explaining how to setup cron in a Docker container here : https://esc.sh/blog/cron-jobs-in-docker/

Answer 4

I know this is an old question but I found a fix to this on Debian and it solved my problem. Cron pam auth with uid was breaking my cron from being able to run.

RUN sed -i '/session    required     pam_loginuid.so/c\#session    required     pam_loginuid.so/' /etc/pam.d/cron