Investigate PostgreSQL shared memory exhaustion after repeated integration test runs

[ValeraFinebits]

Summary

After several integration test runs, BTCPayServer startup begins failing with:

SqlState 53100: could not resize shared memory segment ... No space left on device

Because of this, tests time out before payjoin logic is reached.

Observed in

Example failing test:

• RunTestPaymentIntegrationTests.CreateInvoiceAndPayItThroughRunTestPayment

Failure point:

• PayjoinAccountTestHelper.CreateInitializedTestContextAsync()
• tester.StartAsync().WaitAsync(cancellationToken)

Logs show:

• repeated migration retries
• BTCPayServer.Hosting.MigrationStartupTask: Error while running migration scripts, retrying...
• Npgsql.PostgresException 53100

Why investigate

It is not yet clear whether this is:

• leaked test resources / bad cleanup,
• container or PostgreSQL shared memory exhaustion,
• or something that could also matter in production.

Expected behavior

Repeated test runs should not exhaust resources or break startup.

Acceptance criteria

• identify root cause;
• determine whether this is test-only or production-relevant;
• define a reliable fix or mitigation;
• confirm repeated test runs no longer trigger the issue.