Add ability to set user on Tasks

[tcdowney]

Issue: #4372

• I have reviewed the contributing guide

• I have viewed, signed, and submitted the Contributor License Agreement

• I have made this pull request to the main branch

• I have run all the unit tests using bundle exec rake

• [ in progress ] I have run CF Acceptance Tests

[spgreenberg]

@tcdowney @Gerg We are seeing errors querying the task API for existing applications once this change is applied. Tasks still execute successfully but querying the task API fails. The error message is vague:

cf tasks <app>

Unexpected Response
Response Code: 500
Request ID:    d346f3ca-6195-...
Code: 0, Title: , Detail: {"errors":[{"title":"UnknownError","detail":"An unknown error occurred.","code":10001}]}
FAILED

This only fails for apps that existed before the update and had tasks ran. I suspect this points to a nil/null issue related the missing user in the task records created before the update.

Please let me know if there is additional info I can provide to help track this down.

[Gerg]

The underlying error should be available in the cloud controller logs.

I wasn't able to reproduce this by deploying an older version, running a task, deploying a new version, and then listing tasks. I attempted to reproduce on an environment using MySQL; are you using Postgres?

[spgreenberg]

Thanks, I was able to find some logs. This impacted multiple customers before we rolled back.

{"timestamp":"2025-07-17T12:20:50.127204482Z","message":"Completed 500 vcap-request-id: d346f3ca-6195...","log_level":"info","source":"cc.api","data":{"request_guid":"d346f3ca-6195...","user_guid":"...","b3_trace_id":"d346f3ca...","b3_span_id":"bc24...","status_code":500,"time_taken_in_ms":12,"request_method":"GET","request_fullpath":"/v3/apps/bb94003e.../tasks?per_page=5000"},"thread_id":16200,"fiber_id":16220,"process_id":12,"file":"/var/vcap/data/packages/cloud_controller_ng/82618a.../cloud_controller_ng/lib/cloud_controller/logs/request_logs.rb","lineno":33,"method":"complete_request"}

Yes, we are using postgres.

[spgreenberg]

We have been trying to recreate this issue in our dev environment and have not been able to yet. Oddly we saw this in our dev and prod environments at the same time. We are still working on it.

[Gerg]

I tried to reproduce with Postgres, and it still worked fine. My guess is it's something specific to those apps/tasks.

There should be a stack trace in the logs coming from here.

It should look something like:

{"timestamp":"...","message":"Request failed: 500: {\"errors\"=\u003e[{\"title\"=\u003e\"UnknownError\", \"detail\"=\u003e\"An unknown error occurred.\", \"code\"=\u003e10001, \"test_mode_info\"=\u003e{\"detail\"=\u003e\"
<actual error message>\", \"title\"=\u003e\"<actual error type>\", \"backtrace\"=\u003e[<stack trace>...

[cweibel]

Having a heck of a time reproducing the problem, but DID see what Steve saw. Unfortunately the two apps that had the problem have either been deleted or redeployed which in each case made the problem go away.

There are a few logs the next day with PG::UndefinedColumn: ERROR: column processes.user does not exist but those I attribute to the short period of time (2 minutes?) between when I rolled back the schema changes manually and pinned back and started the deployment in Concourse.

If we find anything more definitive we'll drop an update here.

[spgreenberg]

@cweibel Did notice missing droplets around the same time. Perhaps this is our root issue: #4467.

One of the apps should not have had its droplets purged though. It could have, but this would be impossible for us to validate at this point.