Spurious error on DB setup

DB setup has been observed to fail in the hosted service with the following error logs:

2022-03-16T19:29:16.264973019Z stderr F thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: JoinError::Panic(...)', /graph-node/store/postgres/src/connection_pool.rs:513:10
2022-03-16 19:29:16	
2022-03-16T19:29:16.264648288Z stderr F Mar 16 19:29:16.264 CRIT migrations failed, error: store error: failed to find plan for subquery ss, pool: main, shard: blocks_a, component: ConnectionPool
2022-03-16 19:29:16	
2022-03-16T19:29:16.264640948Z stderr F note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
2022-03-16 19:29:16	
2022-03-16T19:29:16.264553508Z stderr F thread 'tokio-runtime-worker' panicked at 'migrations failed: store error: failed to find plan for subquery ss', store/postgres/src/connection_pool.rs:987:13

This causes graph node to crash, until it the setup eventually succeeds for all shards. It's been observed on both PG 13 and PG 14 shards.

Commencez dans store/postgres/src/connection_pool.rs, au niveau des échecs signalés autour des lignes 513 et 987, et suivez la manière dont les migrations gèrent le plan de sous-requête ayant échoué pour le shard blocks_a. Reproduisez la configuration de la DB sur des shards PostgreSQL 13 et 14, en utilisant les logs fournis et RUST_BACKTRACE=1 pour obtenir davantage de détails. C’est terminé lorsque la configuration ne plante plus face à cet échec transitoire et se termine correctement pour tous les shards.