"No URL for result:" Error with Rails Analyzer

[davereber]

When I try to run the rails analyzer I always get an error for all results stating "No URL for result: ". This occurs for results that I add manually and those that are added via the Github sync. I have my github and tmp download variables set i /config/initializers/scumblr.rb with my GitHub API key and /tmp/. I have been having no luck finding this error documented anywhere.

[ahoernecke]

Hi @davereber,

It looks like the rails analyzer was expecting a field to be added that wasn't be added through the github sync. Can you re-run your github sync then run the rails analyzer again and see if that helps?

[davereber]

Hi @ahoernecke

I just re-ran the github sync. I also ensured that an api.github.com was configured as the api endpoint. I am still getting the same error message. Is there a way to view more detailed logging via the rails console? I am new to using Scumblr.

I did confirm that a URL field exists...but I am assuming that it is looking for something else.

Thanks.

[ahoernecke]

Did you update your code to the latest from master? (sorry forgot to tell you to do so in my original response)?

[davereber]

I did the install (following the wiki instructions) yesterday.

[ahoernecke]

Got it, I pushed a fix for your issue this morning. You'll need to update to the latest code. From your scumblr folder:

git pull origin master

Restart your services (including sidekiq) and rerun the sync.

[davereber]

Looks like it is running :) I will let you know when it finishes the job if there are any errors.

Thank you!

[ahoernecke]

Sounds good, thanks!

[davereber]

Now it seems to execute the Rails Analyzer task, but the job never completes. When I look up in the running tasks it says "No Tasks Running" and I do not see a completed entry in the events. When I try to launch another task nothing will start until I restart the server. I let it run for 4x the amount of time that it took me to manually run brakeman against the same repo.

[ahoernecke]

Hi @davereber,

Any chance you can look at the memory on your system for the time period the scan was running and see if it was being exhausted? We're actively working on fixing some memory issues that would cause this behavior that seemed to be triggered by a combination some inefficient code on our end, a possible memory leak in brakeman, and (at least for us) a couple repos that brakeman really didn't like and would never finished parsing. Your comment about new tasks not launching makes me suspect you may be running into something similar.

We're actively committing to the async-refactor branch working on fixing these memory issues. Obviously it hasn't been promoted to master yet so YMMV, but so far we've seen some pretty significant improvements.