Checking for creation of logfile, without CPU spin, but bail if too many errors

[sychan]

Commit 4c86ed7 fixed the CPU spinning issue, but it also causes dockerize to bail out immediately if a logfile hasn't been created yet by the program to be run when the tail thread goes to look for it. This race condition has broken a lot of my images built using dockerize, and I imagine that it may effect others.

This PR uses time.Sleep() to have the tail thread sleep 2 seconds before retrying, and allows a certain number of retries before bailing out. The number of retries is set for 30, which is gives about a minute for a program to initialize and get its logs going - I'm happy to set it lower if a minute seems too long.

Tailing named pipes is still supported, but the code assumes that if this is a named pipe, mkfifo has already been called before dockerize gets executed.

[jwilder]

Thanks @sychan!

[apollo13]

Hi @sychan. when I run your changes from master (since this is not in a released version yet); I do get:

2018/12/30 13:22:20 Warning: unable to stat /var/log/nginx/error.log: stat /var/log/nginx/error.log: no such file or directory
2018/12/30 13:22:20 Warning: unable to stat /var/log/nginx/access.log: stat /var/log/nginx/access.log: no such file or directory
2018/12/30 13:22:27 Warning: unable to tail /var/log/nginx/error.log: tomb: still alive
2018/12/30 13:22:29 Warning: unable to tail /var/log/nginx/error.log: tomb: still alive
2018/12/30 13:22:31 Warning: unable to tail /var/log/nginx/error.log: tomb: still alive
2018/12/30 13:22:36 Warning: unable to tail /var/log/nginx/error.log: tomb: still alive
2018/12/30 13:22:40 Warning: unable to tail /var/log/nginx/error.log: tomb: still alive
2018/12/30 13:22:42 Warning: unable to tail /var/log/nginx/error.log: tomb: still alive

The first two lines do make sense, the logfiles are not there yet, but it looks as if it runs into another error afterwards. For reference I am using this dockerfile https://github.com/jwilder/dockerize/blob/b6062ae239adee8b2e7fa464ceaea1721fdd1ccb/examples/nginx/Dockerfile with /var/log/nginx/* cleared… do you have any ideas?

[apollo13]

Or even another issue, I use this dockerfile:

FROM ubuntu:18.04
RUN apt-get update && apt-get install -y nginx
RUN echo "daemon off;" >> /etc/nginx/nginx.conf
RUN rm /var/log/nginx/*
EXPOSE 80
CMD ["dockerize", "-stdout", "/var/log/nginx/access.log", "-stderr", "/var/log/nginx/error.log", "nginx"]

and run with:

docker run --rm -v /home/florian/go/bin/:/usr/local/bin -p 8080:80 test2

where I map my go path in so I do get the local dockerize version, I get no output at all aside from:

2018/12/30 13:29:48 Warning: unable to stat /var/log/nginx/error.log: stat /var/log/nginx/error.log: no such file or directory
2018/12/30 13:29:48 Warning: unable to stat /var/log/nginx/access.log: stat /var/log/nginx/access.log: no such file or directory

once.