Cron failing in TravisCI

[dougbtv]

Looks like the builds have been failing in Travis. Apparently at the point where it looks to see if the cron job[s] ran OK.

Mostly a note for myself to go in and debug what's causing the CI failure (or anyone else who wants to take a look!).

[lmangani]

Have you noticed if this is for multi-container or all-in-one? or both?

[dougbtv]

It's for multi-container, I don't think I have any end-to-end testing of the AIO container, just tests if the build completes for it. But e2e tests for the multi-container actually spins up the containers, then checks that cron ran and created the tables, further from there gens some traffic with hepgen and looks for a result in mysql.

[arainero]

@dougbtv

I think I am suffering from this issue or something similar to it. I notice that after running Homer for more than one day and less than 4 nothing gets recorded anymore. I don't have an exact time frame at the moment.

The temporary solution was to do a docker-compose down and docker-compose up.

After googling around my issue seems related to this.

[lmangani]

@dougbtv hey buddy there was an attempt to remove cron from the mix here, up for a look and comments: https://github.com/lmangani/homer-nocron

Thanks!

[dougbtv]

I will take a @lmangani -- quick glance looks pretty great. I'll give it a whirl