Skip to content

Migrate origin_visit_status records to add the type value

Once the new scheduler pattern is deployed. We'll need to align the origin_visit_status data records to actually have their type filled in.

That means:

  • modify the schema to add the visit_type column
  • sql file to migrate the data from origin_visit to origin_visit_status
  • Check impacts on storage clients (i.e. loaders) (already dealt with indeed)

Data to migrate 1239705800 rows out of 1259171054:

Production status (mirror):

#+begin_src sql
softwareheritage=> select now(), count(*) from origin_visit_status;
              now              |   count
-------------------------------+------------
 2021-02-02 09:47:15.904791+00 | 1259171054
(1 row)

softwareheritage=> select now(), count(*) from origin_visit_status where type is null;
              now              |   count
-------------------------------+------------
 2021-02-02 09:53:08.903614+00 | 1239705800
(1 row)
#+end_src

Deliverables migrate both:

  • staging db (as it will be used to exercise the migration script and it must be up-to-date as well)
  • checks
  • production db (details below)
  • #2968 (closed): checks

Migrated from T2968 (view on Phabricator)

Edited by Antoine R. Dumont