Error in google cloud run connecting with cloud SQL

The error message i get when running in docker with cloud run:

"\nInvalid prisma.() invocation in\n/usr/src/app/src/federations/b2bike/entities/service-pass/resolvers.js:6:76\n\n\n\nReason: Error occurred during query execution:\nConnectorError(ConnectorError { user_facing_error: None, kind: ConnectionError(Timeout\n\nstack backtrace:\n 0: failure::backtrace::internal::InternalBacktrace::new\n 1: failure::backtrace::Backtrace::new\n 2: sql_query_connector::error::SqlError::into_connector_error\n 3: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 4: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 5: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 6: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 7: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 8: std::future::poll_with_tls_context\n 9: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 10: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 11: std::panicking::try::do_call\n 12: __rust_maybe_catch_panic\n at src/libpanic_unwind/lib.rs:78\n 13: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 14: <std::future::GenFuture<T> as core::future::future::Future>::poll\n 15: hyper::proto::h1::dispatch::Dispatcher<D,Bs,I,T>::poll_catch\n 16: <hyper::server::conn::spawn_all::NewSvcTask<I,N,S,E,W> as core::future::future::Future>::poll\n 17: tokio::task::core::Core<T>::poll\n 18: std::panicking::try::do_call\n 19: __rust_maybe_catch_panic\n at src/libpanic_unwind/lib.rs:78\n 20: tokio::task::harness::Harness<T,S>::poll\n 21: tokio::runtime::thread_pool::worker::GenerationGuard::run_task\n 22: tokio::runtime::thread_pool::worker::GenerationGuard::run\n 23: std::thread::local::LocalKey<T>::with\n 24: tokio::runtime::thread_pool::worker::Worker::run\n 25: <std::panic::AssertUnwindSafe<F> as core::ops::function::FnOnce<()>>::call_once\n 26: std::panicking::try::do_call\n 27: __rust_maybe_catch_panic\n at src/libpanic_unwind/lib.rs:78\n 28: tokio::task::harness::Harness<T,S>::poll\n 29: tokio::runtime::blocking::pool::Inner::run\n 30: tokio::runtime::context::enter\n 31: std::sys_common::backtrace::__rust_begin_short_backtrace\n 32: std::panicking::try::do_call\n 33: __rust_maybe_catch_panic\n at src/libpanic_unwind/lib.rs:78\n 34: core::ops::function::FnOnce::call_once{{vtable.shim}}\n 35: <alloc::boxed::Box<F> as core::ops::function::FnOnce<A>>::call_once\n at /rustc/73528e339aae0f17a15ffa49a8ac608f50c6cf14/src/liballoc/boxed.rs:942\n 36: <alloc::boxed::Box<F> as core::ops::function::FnOnce<A>>::call_once\n at /rustc/73528e339aae0f17a15ffa49a8ac608f50c6cf14/src/liballoc/boxed.rs:942\n std::sys_common::thread::start_thread\n at src/libstd/sys_common/thread.rs:13\n std::sys::unix::thread::Thread::new::thread_start\n at src/libstd/sys/unix/thread.rs:79\n 37: start_thread\n 38: __clone\n) })\n"

[dkozma]

A Cloud Run-compatible socketpath can be provided to Prisma, for instance:

postgres://USER:PASSWORD@localhost/DBNAME?host=/cloudsql/CLOUD_SQL_INSTANCE_CONNECTION_NAME

[diegocouto]

Hi everyone,

Updated message (June 04, 2020):

I was able to solve this issue using ?socket= instead of ?host=. This was suggested in this thread before but, as the documentation was updated after it’s been discussed here, I was under the impression that this had been updated as well.

Original Message

Were any of you able to set up a socket connection when using MySQL as the provider? I’ve been trying to do so but I keep seeing the error:

"extensions": {
  "code": "INTERNAL_SERVER_ERROR",
  "exception": {
    "code": "P1001",
    "meta": {
      "database_host": "localhost",
      "database_port": 3306
    },
    "stacktrace": [
      "Error: ",
      "Invalid `prisma.user.findOne()` invocation in",
      "/usr/src/app/build/controllers/UserController.js:113:68",
      "",
      "",
      "  Can't reach database server at `localhost`:`3306`",
      "",
      "Please make sure your database server is running at `127.0.0.1`:`3306`.",
      "    at PrismaClientFetcher.request (/usr/src/app/node_modules/@prisma/client/runtime/index.js:1:148505)",

In order to limit potential issues, I’m hardcoding the data source URL while testing this as such:

datasource db {
  provider = "mysql"
  url      = "mysql://user:psswd@localhost/db?host=/cloudsql/project:region:db"
}

To make sure the issue was not related to some missing permissions between Cloud Run and Cloud SQL, I’ve created a Knex client trying to connect to the same socket and it does connect successfully.

I’m not sure whether this could be an indicator or not but it’s weird that the error message never mentions the socket, only the IP:PORT, making it look like it’s always trying a TCP connection only.

Any thoughts? ☺️