Running multiple findUnique's in parallel causes both to return null

Bug description

When running two of the same findUnique queries in parallel, both return null where they should return the same document.

How to reproduce

1. Create a new Prisma project with a PostgreSQL database.
2. Add a table with a @@unique attribute

datasource db {
  provider = "postgresql"
  url      = "URL"
}

generator client {
  provider = "prisma-client-js"
}

model Property {
  id           String         @id @default(cuid())
  Availability Availability[]
}

model Availability {
  date          DateTime
  property      Property              @relation(fields: [propertyId], references: [id])
  propertyId    String
  status        AvailabilityStatus    @default(AVAILABLE)

  @@unique([propertyId, date, status])
}

enum AvailabilityStatus {
  AVAILABLE
}

1. Create a new set of documents in the table

const property = await prisma.property.create({});

const targetDate = new Date('2020-12-03T00:00:00.000Z');

await prisma.availability.create({
    data: {
      property: { connect: {id: property.id }},
      status: AvailabilityStatus.AVAILABLE,
      date: targetDate,
    },
  });

1. Query the document twice in parallel

const firstUnique = prisma.availability.findUnique({
    where: {
      propertyId_date_status: {
        propertyId: property.id,
        status: AvailabilityStatus.AVAILABLE,
        date: targetDate,
      }
    },
  });

  const secondUnique = prisma.availability.findUnique({
    where: {
      propertyId_date_status: {
        propertyId: property.id,
        status: AvailabilityStatus.AVAILABLE,
        date: targetDate,
      }
    },
  });

  const res = await Promise.all([firstUnique, secondUnique])

  console.log(res); <-- [null, null]

Expected behavior

Expected to return the same document twice - a warning would also be super helpful if this happens.

Prisma information

Environment & setup

• OS: Mac
• Database: PostgreSQL
• Node.js version: 12.18.0
• Prisma version: 2.12.1

@prisma/cli          : 2.12.1
@prisma/client       : 2.12.1
Current platform     : darwin
Query Engine         : query-engine cf0680a1bfe8d5e743dc659cc7f08009f9587d58 (at node_modules/@prisma/engines/query-engine-darwin)
Migration Engine     : migration-engine-cli cf0680a1bfe8d5e743dc659cc7f08009f9587d58 (at node_modules/@prisma/engines/migration-engine-darwin)
Introspection Engine : introspection-core cf0680a1bfe8d5e743dc659cc7f08009f9587d58 (at node_modules/@prisma/engines/introspection-engine-darwin)
Format Binary        : prisma-fmt cf0680a1bfe8d5e743dc659cc7f08009f9587d58 (at node_modules/@prisma/engines/prisma-fmt-darwin)
Studio               : 0.322.0

[nrxus]

We just run into this behavior and it was incredibly surprising. The most disappointing part is how how silent this error is. There is no warning in the prisma logs, nothing documented about using composite keys with findUnique not working when doing things in parallel. While I understand that this bug may be hard to resolve, at the very least some loud logging or just a panic so it can be caught quickly would be the least I would expect from a tool that should be prod-ready.

EDIT: In retrospect my comment comes accross a little too aggressive/accusatory, I apologize for my tone. Commenting after spending a long time debugging got the better of me.

[nrxus]

I do see it fixed in 2.15! Thank you so much!

I should have tried updating my prisma before commenting, I was on 2.14 and since I didn’t see anything in the changelog for 2.15 about this I assumed it was still an issue 🤦🏽