scrubber: Investigate the apparent lock (staging)

Plenty of apparently stuck insert queries in the db [1] with a db still empty [2] (out of [3]) So something must be wrong somewhere.

Let's investigate and fix that.

• [1]

1022114 swh-scrubber                       swh-scrubber     127.0.0.1/32    0.0  0.0       0B       0B     503 h    transactionid    N             active   INSERT INTO datastore (package, class, instance) VALUES ('storage', 'postgresql',
'user=guest password=xxx dbname=swh host=db1.internal.staging.swh.network port=5432') ON CONFLICT DO NOTHING RETURNING id

And nothing actually written in the scrubber db as of yet since its deployment around 3 weeks ago.

15:24:26 swh-scrubber@db1:5432=> select * from corrupt_object ;
+----+-----------+--------+------------------+
| id | datastore | object | first_occurrence |
+----+-----------+--------+------------------+
+----+-----------+--------+------------------+
(0 rows)

Time: 7.313 ms
15:29:39 swh-scrubber@db1:5432=> select * from datastore;
+----+---------+-------+----------+
| id | package | class | instance |
+----+---------+-------+----------+
+----+---------+-------+----------+
(0 rows)

Time: 13.705 ms
15:29:43 swh-scrubber@db1:5432=> select * from object_origin ;
+-----------+------------+--------------+
| object_id | origin_url | last_attempt |
+-----------+------------+--------------+
+-----------+------------+--------------+
(0 rows)

Time: 5.291 ms

• [3]

root@db1:~# sudo -i -u postgres pg_activity -p 5433 dbname=swh-scrubber

---

Migrated from T4228 (view on Phabricator)