Unstuck failing loaders in staging

Loaders (mostly package ones) currently fail to load [1]

• [1] https://sentry.softwareheritage.org/share/issue/4eb69ea2cc314c4ab811809ff60c358f/

---

Migrated from T2428 (view on Phabricator)

mentioned in merge request !88 (closed)

assigned to @ardumont

added Core Loader Staging environment priority:Normal labels

changed the description

mentioned in commit 385843d8

In the end, it's not necessarily data [1] which is obsolete. It's more having partial production dump mounted on staging. We have dangling data.

For example, recently a dumps with origin/origin-visits was mounted. This means we reference snapshot which do not exist right now.

That's making the npm/pypi/etc... loaders blow up. I'm fixing that in the loader for now. It's a technical path which could happen in production anyway (because the storage api used in the loader allows it)...

• [1] I was referring to scheduling data

changed title from refresh staging loading data in scheduler to Unstuck failing loaders in staging

changed the description

added state:wip label

Deployed.

removed state:wip label

closed