Bail if upstream URL is malformed

[amartin96]

In at least one case (postgres missing db name), a malformed upstream URL results in errors that could be cleaner (indefinitely repeating replicator errors).

We should make this case a permanent failure with a user-friendly message, and investigate/fix possible others.

_{From SyncLinear.com | REA-2917}

[amartin96]

Postgres can be configured to allow passwordless login as well. It does, however, always require a username, as far as I've read.

[amartin96]

In MySQL, you can create a superuser with no username or password:

CREATE USER ''@'%' IDENTIFIED BY '';
GRANT ALL PRIVILEGES ON *.* TO ''@'%';

When such a user exists, ReadySet will connect and replicate from a URL of the form:

mysql://<host>:<port>/<db>

So we don't want to require username + password for MySQL deployments.

[amartin96]

Postgres username missing:

Error { kind: Db, cause: Some(DbError { severity: "FATAL", parsed_severity: Some(Fatal), code: SqlState(E28000), message: "no PostgreSQL user name specified in startup packet", detail: None, hint: None, position: None, where_: None, schema: None, table: None, column: None, datatype: None, constraint: None, file: Some("postmaster.c"), line: Some(2268), routine: Some("ProcessStartupPacket") }) }

Postgres password missing:

Error { kind: Config, cause: Some("password missing") }

Postgres bad username/password:

Error { kind: Db, cause: Some(DbError { severity: "FATAL", parsed_severity: Some(Fatal), code: SqlState(E28P01), message: "password authentication failed for user \"postgress\"", detail: None, hint: None, position: None, where_: None, schema: None, table: None, column: None, datatype: None, constraint: None, file: Some("auth.c"), line: Some(335), routine: Some("auth_failed") }) }

Conclusions:

It might be possible to differentiate between recoverable and unrecoverable errors based on this information, but it would involve matching on strings and/or opaque error codes, which is brittle. We should just check for these fields manually.

[amartin96]

Blocked because I think the most robust way to handle these errors is to treat them as unrecoverable where they are now, rather than add logic elsewhere to try and catch them early.

[amartin96]

Omitting the username or password also leads to a replicator failure loop.

[amartin96]

https://gerrit.readyset.name/c/readyset/+/5197

[amartin96]

Some discussion on whether readyset might ever want to deal with postgres upstream URLs that don't contain a database name:

https://readyset-workspace.slack.com/archives/C01HF1WTP6Y/p1686944393015599