Travis CI runs out of memory since latest docker image update (17-11-2017)

[jfloff]

Travis CI runs out of memory since latest docker image update (17-11-2017)

[jfloff]

That's interesting ... I've indeed triggered a new image build due to this issue #15, but since I didn't change anything on the image I wasn't expecting nothing to change.

Well lets try narrow it down, which image tag are you using?

[snoepkast]

We're using 2.7, if you want to take a look at our dockerfile, it's here. The Ci we're using is Travis (also contacted them about it), builds can be found here.

After local tests it doesn't seem like the memory has actually increased though, we tested by building using the cached image and the new one while monitoring with docker stats.

Maybe you have a clue what goes wrong? The builds just stall...

[jfloff]

I've cloned your repo, and I'm trying to build the image locally to see if I see any errors while building. I'll be back when its done

[jfloff]

I've builded the image without any problem. I've looked through your Ci logs and it seems that its hanging perhaps on the wait-for-it script. I'm running the docker-compose now to do a couple of tests.

[snoepkast]

Ah yeah we've noticed that now too, it appears to be the exit statement or something (before we updated our dependencies it stalled in the middle of the pip install gevent so maybe that has exit too). Trying it now with the new docker healthchecks.

[jfloff]

I was able to run the dockerfile and run the tests directly (python manage.py test --with-coverage --cover-erase --cover-package=lily), so I would say something is wrong with the wait-for-it.

I went to its repo, and found out this commit from 4 months ago that has a bug fix for when you pass parameters. I've checked and you indeed pass parameters wait-for-it.sh -b db:5432 and your version of the script is before that commit. Maybe you could test updating the wait-for-it script and rerun the build?

[snoepkast]

Yeah, thanks for your help!