r2dbc-postgresql support 40001 error code (serialization_failure)?Currently, when db returns 40001,
io.r2dbc.postgresql.ExceptionFactory puts this error in io.r2dbc.postgresql.PostgresqlTransientExceptionand then in
org.springframework.r2dbc.connection.ConnectionFactoryUtils it is wrapped in org.springframework.r2dbc.UncategorizedR2dbcException.
Hi everyone. We are currently facing a strange issue were queries are not executed
Details: We have a reactive web application which is secured using spring security and so we wrote a ReactiveUserDetailsService which loads our user using Spring-R2DBC Repositories in two steps:
- Load the user
- Load the authorities if the user was found in the database
Code looks something like this:@Override public Mono<UserDetails> findByUsername(final String username) { final Mono<UserProjection> userMono = userRepository.findProjectionByUsername(username); return userMono .flatMap(u -> { return authorityRepository.findByUsername(u.getUsername()) .map(a -> new SimpleGrantedAuthority(a.getAuthority())) .collectList() .flatMap(l -> Mono.just( User.withUsername(u.getUsername()).password(u.getPassword()) .authorities(l).disabled(!u.isActive()).build())); }); }
Now sometimes it happens that this method is executed (a breakpoint in there is hit), but there is not a single Query executed (according to the logs with logging.level.org.springframework.r2dbc = DEBUG)
So even the userRepository.findProjectionByUsername() call doesn't do anything
Anyone an idea what happens there?
@mp911de Regarding the issue:
My thoughts are that command COPY TO STDOUT should be suported the same way the DataRow is suported in src/main/java/io/r2dbc/postgresql/PostgresqlResult.java because it is result and official documentations says that:
The backend sends a CopyOutResponse message to the frontend, followed by zero or more CopyData messages (always one per row), followed by CopyDone.
Which tells me that PostgresqlRow::toRow should maybe be overloaded for CopyDatatype of message also. And as per documentation it should be expected that driver is capable of delivering the COPY TO STDOUT result to the client.
When STDIN or STDOUT is specified, data is transmitted via the connection between the client and the server.
For example:
if (message instanceof CopyData) {
PostgresqlRow row = PostgresqlRow.toRow(this.context, (CopyData) message, this.rowDescription);
sink.next(f.apply(row, this.metadata));
}Now this should probably for start be supported in simple query protocol as I don't see any obvious use for exteneded one.
For the COPY FROM STDIN I guess we should be somehwere close to client folder as per documentation we should sent data as multiple CopyData messages. Now, STDIN does not need to be real STDIN in a sense, it could be file, or java stream, some array maybe etc. so that support should go to the cliend in some sense.
Now it's just high overview of the idea, I just wan't to get your opinion, to see is it the right direction and provide more detailed overview soon. And I know I took some time, but I'm little junior to reactive Java, this driver and life happens so :D
Thanks :bow:
map because it already accepts a BiFunction. Maybe we should even consider an extended API design similar to Mutiny (on().row().map(…), on().copy().map()) but that's more food for thought.
Statement API, either withCopyIn or overload execute accepting a publisher of buffers or a CopyIn wrapper.
preparedStatementCacheQueries when using the URL-based configuration mechanism or the preparedStatementCacheQueries() method on the config builder @IvanGTrendafilov
9 replies
Please help with investigation of it. I understand that the test looks like a bit far-fetched, but it models our case we faced during development of billing application with CockroachDB as its database.
Hi, thanks for the good work.
I would like to use r2dbc-proxy together with r2dbc-pool. Here is what I have so far:
final ConnectionFactory connectionFactory = ConnectionFactories.get(ConnectionFactoryOptions.builder()
.option(DRIVER, "pool")
.option(PROTOCOL, "postgresql")
.option(PORT, postgreSqlReactiveConfig.getPort())
.option(HOST, postgreSqlReactiveConfig.getHost())
.option(USER, postgreSqlReactiveConfig.getUsername())
.option(PASSWORD, postgreSqlReactiveConfig.getPassword())
.option(DATABASE, postgreSqlReactiveConfig.getDatabase())
.option(MAX_SIZE, connectionPoolConfig.getMaxSize())
.option(PostgreSqlReactiveConfig.PREPARED_STATEMENT_CACHE_QUERIES_OPTION, 0)
.build());How this can be combined with
ConnectionFactory connectionFactory = ConnectionFactories.get(ConnectionFactoryOptions.builder()
.option(DRIVER, "proxy")
.option(PROTOCOL, "postgresql")
.build());r2dbc:proxy:pool:postgresql://host:port/?maxSize=…& preparedStatementCacheQueries =… makes clear where this comes from.
10:14:36.179 [reactor-tcp-nio-2] WARN io.r2dbc.postgresql.client.ReactorNettyClient - Error: SEVERITY_LOCALIZED=ERROR, SEVERITY_NON_LOCALIZED=ERROR, CODE=23505, MESSAGE=duplicate key value violates unique constraint "users_email_key", DETAIL=Key (email)=(email) already exists., SCHEMA_NAME=public, TABLE_NAME=users, CONSTRAINT_NAME=users_email_key, FILE=nbtinsert.c, LINE=656, ROUTINE=_bt_check_unique
hi! i have a question about r2bdc-pool which i couldn't figure out since i am not so great at reading reactive code. what happens when obtaining a connection from the pool and the connection validation fails? (for example, because the backend server was restarted.) does it try to create a new connection?
i'm looking at https://github.com/r2dbc/r2dbc-pool/blob/a58db10f548997eb1ea6074fe5d5ba3483ca8cf9/src/main/java/io/r2dbc/pool/ConnectionPool.java#L109
I've applied the following r2dbc-proxy settings:
List<ProxyExecutionListener> proxyExecutionListeners = new ArrayList<>();
proxyExecutionListeners.add(tracingExecutionListener);
final ConnectionFactory connectionFactory = ConnectionFactories.get(ConnectionFactoryOptions.builder()
.option(DRIVER, "proxy")
.option(PROTOCOL, "pool:postgresql")
.option(PORT, postgreSqlReactiveConfig.getPort())
.option(HOST, postgreSqlReactiveConfig.getHost())
.option(USER, postgreSqlReactiveConfig.getUsername())
.option(PASSWORD, postgreSqlReactiveConfig.getPassword())
.option(DATABASE, postgreSqlReactiveConfig.getDatabase())
.option(MAX_SIZE, connectionPoolConfig.getMaxSize())
.option(PostgreSqlReactiveConfig.PREPARED_STATEMENT_CACHE_QUERIES_OPTION, 0)
.option(PostgreSqlReactiveConfig.PROXY_LISTENER_OPTION, proxyExecutionListeners)
.build());
final ConnectionPoolConfiguration configuration = ConnectionPoolConfiguration.builder(connectionFactory)
.initialSize(connectionPoolConfig.getInitialSize())
.maxSize(connectionPoolConfig.getMaxSize())
.maxIdleTime(Duration.ofMillis(connectionPoolConfig.getMaxIdleTime()))
.maxCreateConnectionTime(Duration.ofSeconds(connectionPoolConfig.getMaxCreateConnectionTime()))
.acquireRetry(connectionPoolConfig.getAcquireRetry())
.maxAcquireTime(Duration.ofMillis(connectionPoolConfig.getMaxAcquireTime()))
.build();
return new ConnectionPool(configuration);Everything works fine when requests are done to our WebFlux controllers, but when we have a thrift request we get this:
org.springframework.dao.DataAccessResourceFailureException: Failed to obtain R2DBC Connection; nested exception is java.lang.IllegalArgumentException: interface io.r2dbc.spi.Connection is not visible from class loader@mp911de any ideas?
I'm not sure but the problem might come from
return (ConnectionFactory) Proxy.newProxyInstance(Thread.currentThread().getContextClassLoader(),
new Class<?>[]{ConnectionFactory.class, Wrapped.class, ProxyConfigHolder.class}, invocationHandler);based on https://stackoverflow.com/a/13379769
sun.misc.Launcher$AppClassLoader@18b4aac2
ClassLoader seems the better choice.