Code review comment for lp:~edwin-grubbs/launchpad/bug-615654-registry-jobqueue-schema

I'm happy with the names. These relations have so far often been named by the tables they are joining (BranchJob, ArchiveJob) as with the job_type column, they can be used for several different types of task. Although we are not consistent - ApportJob should be BlobJob for instance.

We want a unique database user per script. If you have one script processing multiple types of jobs, that is one database user.

We will need some indexes in case these tables grow:

CREATE INDEX PersonTransferJob__minor_person__idx ON PersonTransferJob(minor_person);
CREATE INDEX PersonTransferJob__major_person__idx ON PersonTransferJob(major_person);
CREATE INDEX QuestionJob__question__idx ON QuestionJob(question);