Infinite error loop on read time out.

[infectedrhytms]

So I figured why messageReceived wasn't called for my
RequestReadTimeoutHandler, because in the pipeline it was placed after my
IntegerHeaderHandler... But I still get some issues...

I'll explain my test case... Here my pipeline described...

1- RequestReadTimeHandler 20 seconds
2- MyIntegerHeaderHandler
3- Netty's StringDecoder
4- Netty's StringEncoder
5- OrderedMemoryAwareThreadExecutor
6- MyStringToObjectEncoderDecoder
7- My BusinessHandler


Situation 1: Client sends correct message with RequestReadTimeoutHandler
enabled...

I setup my stress 50 requests a second...

Netty starts throwing read timeout errors for nothing.... And resources
start being used up. I have an object in my business handler that uses
awt.QueueEvent... and it trhows stack overflow errors. This only seems to
happens when the read time out handler is enabled... Also
RequestReadTimeoiutHandler throws closed socket excpetions which also makes
the CPU skyrocket 100%, but eventually everything is freed...


Situation 2: Client sends correct message RequestTimeOuthandler not added to
pipeline. No issues at all everything works perfect! So what's the issue?

Also another question lets say I don't implement the read time out! What
happens hypothetically if a client connects and doesn' t send the right
amount of bytes and doesn't close the connection either, will Netty just
discard everything quietly? Just trying to foresee what if a bad client
connects and does bad things out of my control...








-- 
View this message in context: http://netty-forums-and-mailing-lists.685743.n2.nabble.com/Infinite-error-loop-on-read-time-out-tp5008996p5016953.html
Sent from the Netty User Group mailing list archive at Nabble.com.