Console warnings in "hello world" app.

[hsyed]

Decided to give the driver a go. But getting warnings. There is a bazillion of these warnings when some actors go down in an akka app, and they show up in this hello world chunk of code as well.

Test extends App {
    println("starting")
    val configuration = URLParser.parse("jdbc:postgresql://localhost/crawler?user=hassan&password=")
    val connection = new PostgreSQLConnection(configuration)
    Await.result(connection.connect, 5.seconds)

    println("connect")

    val fut =  connection.sendQuery("select 229;")
    val actualResult = Await.result(fut,5 seconds)

    println(actualResult.rows.map(_.head).map(_.apply(0)).get)

    val res = Await.result(connection.disconnect,5.seconds)
    println("disconnect")
    Thread.sleep(5000)
}

Warning :

18:59:21.236 [pool-6-thread-1] WARN  io.netty.channel.nio.NioEventLoop - Selector.select() returned prematurely 512 times in a row; rebuilding selector.
18:59:21.236 [pool-6-thread-1] INFO  io.netty.channel.nio.NioEventLoop - Migrated 0 channel(s) to the new Selector.
[success] Total time: 10 s, completed 08-Dec-2013 18:59:21
18:59:21.242 [pool-6-thread-1] WARN  io.netty.channel.nio.NioEventLoop - Selector.select() returned prematurely 512 times in a row; rebuilding selector.
18:59:21.242 [pool-6-thread-1] INFO  io.netty.channel.nio.NioEventLoop - Migrated 0 channel(s) to the new Selector.

[mauricio]

It's an NIO bug that netty works around, it's just letting you know that it has happened.

[ryanlecompte]

Do you suggest just adjusting the log level for that?

On Dec 8, 2013, at 12:14 PM, Maurício Linhares notifications@github.com wrote:

It's an NIO bug that netty works around, it's just letting you know that it has happened.

— Reply to this email directly or view it on GitHub.

[mauricio]

Is it causing a lot of noise?

I wouldn't do it.

I have it set to debug for the project builds and I don't really see any of that -> https://travis-ci.org/mauricio/postgresql-async/builds/15141939

[hsyed]

Bear in mind that I do not have much experience with this library yet. But, Now that I put a wait on the disconnect and clean up properly it's not a lot of noise. Just 2-3 message pairs per connection close.

However, it was an endless loop when an actor didn't clean up properly. This could explode a log file in a system with a badly written actor. I'd rather have the vm crash than terrorise the log file :D

It is still a very disconcerting message for someone starting with async coding in Java. A low level ASIO event loop failing 512 times in a method called "select" screams "run away" to me. Do you know what is going on underneath ? I would suggest rewriting the error message and make it less dangerous looking :D

[mauricio]

How would you rewrite it?

That's inside Netty itself, not the driver.

Em domingo, 8 de dezembro de 2013, hsyed escreveu:

Bear in mind that I do not have much experience with this library yet. But, Now that I put a wait on the disconnect and clean up properly it's not a lot of noise. Just 2-3 message pairs per connection close.

It is still a very disconcerting message for someone starting with async coding in Java. A low level ASIO event loop failing 512 times in a method called "select" screams "run away" to me. Do you know what is going on underneath ? I might rewrite the error message and make it less dangerous looking :D

— Reply to this email directly or view it on GitHubhttps://github.com/mauricio/postgresql-async/issues/68#issuecomment-30097013 .

Maurício Linhares http://mauricio.github.io/ - http://twitter.com/#!/mauriciojr

[normanmaurer]

@mauricio on a disconnect you close also the EventLoopGroup ?

[mauricio]

@normanmaurer no, there's a single event loop that is kept for all connections as a static variable. Should I create one event loop for every connection?

[mauricio]

@hsyed @ryanlecompte can you guys provide more details about your environment? OS, Kernel, JVM and DB versions?

[hsyed]

Sorry for the delay.

OS X 10.9.1. Scala version 2.10.3 (Java HotSpot(TM) 64-Bit Server VM, Java 1.7.0_45), postgres 9.3.2

[mauricio]

@hsyed do you see this same issue when you run the project tests?

[ryanlecompte]

I've only seen this issue when running using sbt's run-main. If I bundle up a jar and execute that (totally outside of sbt), everything works just fine.

On Mon, Dec 23, 2013 at 12:35 PM, Maurício Linhares < notifications@github.com> wrote:

@hsyed https://github.com/hsyed do you see this same issue when you run the project tests?

— Reply to this email directly or view it on GitHubhttps://github.com/mauricio/postgresql-async/issues/68#issuecomment-31142325 .

[mauricio]

@ryanlecompte can you create a small GH repo with the code you're seeing this?

[ryanlecompte]

I can't right now - about to take off for the holidays! Hopefully @hsyed can, though. It's really just creating a basic connection pool, execute a simple query, then call close()/disconnect() (can't remember name of method exactly). Right after closing, the logging goes crazy for about 6 seconds and then the program finally exits. This is with postgresql, not mysql.

On Mon, Dec 23, 2013 at 12:37 PM, Maurício Linhares < notifications@github.com> wrote:

@ryanlecompte https://github.com/ryanlecompte can you create a small GH repo with the code you're seeing this?

— Reply to this email directly or view it on GitHubhttps://github.com/mauricio/postgresql-async/issues/68#issuecomment-31142464 .