SingleThreadedAsyncObjectPool alternatives

[fwbrasil]

I'm performing some scalability tests using postgresql-async. Currently, the SingleThreadedAsyncObjectPool is the bottleneck. Do you have plans to review this implementation?

I think that a good alternative is to create multiple SingleThreadedAsyncObjectPool that are used according a thread affinity (thead.getId % partitionCount). Something like a PartitionedSingleThreadedAsyncObjectPools.

[mauricio]

Looks great 👍

[fwbrasil]

@mauricio I created an implementation on my fork, but there is something strange.

https://github.com/fwbrasil/postgresql-async/commit/1dad77e2990d5d6f292eb0fe891b2e22e7015043

Looks like Netty is lost when multiple pools are used. The PartitionedSingleThreadedAsyncObjectPoolsSpec fails intermittently with this error:

[ERROR][pool-233-thread-1][2013-11-02 19:42:39,460] Error on connectionjava.io.IOException: Connection reset by peer
at sun.nio.ch.FileDispatcherImpl.read0(Native Method)
at sun.nio.ch.SocketDispatcher.read(SocketDispatcher.java:39)
at sun.nio.ch.IOUtil.readIntoNativeBuffer(IOUtil.java:225)
at sun.nio.ch.IOUtil.read(IOUtil.java:193)
at sun.nio.ch.SocketChannelImpl.read(SocketChannelImpl.java:359)
at io.netty.buffer.UnpooledUnsafeDirectByteBuf.setBytes(UnpooledUnsafeDirectByteBuf.java:436)
at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:870)
at io.netty.channel.socket.nio.NioSocketChannel.doReadBytes(NioSocketChannel.java:208)
at io.netty.channel.nio.AbstractNioByteChannel$NioByteUnsafe.read(AbstractNioByteChannel.java:88)
at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:485)
at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:452)
at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:346)
at io.netty.util.concurrent.SingleThreadEventExecutor$2.run(SingleThreadEventExecutor.java:101)
at java.lang.Thread.run(Thread.java:722)

Could you help with this? My Netty knowledge is very limited.

[mauricio]

Looks like you took too long to write to the server and it gave up cutting this connection off.

This could be a timeout issue (like a small timeout) at the DB end. Gonna try your code later this week and see if I can dig up something.