Domain deletion task is failing

[humitos]

This method,

https://github.com/rtfd/readthedocs.org/blob/df85fefc5f59c7a00bcc1ffe23965efe75e8c895/readthedocs/projects/models.py#L1030-L1034

deletes the Domain from the database and trigger a task called: symlink_domain. This task tries to get the Domain object using the pk from the database and it fails because the object was already removed.

I do see two different approaches here:

1. do not trigger this task and rely on the task to remove orphan symlinks implemented in Task to remove orphan symlinks #3543
2. refactor the task and the helper methods (Symlink.symlink_cnames and Symlink.remove_symlink_cname) to depend only in the Domain.domain string instead of the whole Domain object

Currently, we are depending on 1) but many errors are being generated in our sentry when the Domain is deleted: https://sentry.io/read-the-docs/readthedocs-org/issues/235601094/

I saw this code before, but I didn't noticed that the Domain object was deleted and then tried to get it from the DB inside the task: #3493 (comment)

[dojutsu-user]

@humitos
I would like to work on this issue.
From my understanding,
Domain.domain should be passed instead of Domain.pk in the delete method like this

def delete(self, *args, **kwargs):  # pylint: disable=arguments-differ
        from readthedocs.projects import tasks
        broadcast(type='app', task=tasks.symlink_domain,
                  args=[self.project.pk, self.domain, True],)
        super(Domain, self).delete(*args, **kwargs)

and then correcting the task and other helper methods.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[humitos]

Not stale, we have a PR that deals with this issue.