High CPU usage - may be deadlock

[flashmandv]

When my server gets more clients and respectivelly more data transfer, the CPU is getting high. I made stack dump using "jstack" and here is the log for the problematic thread (I'm using the latest code):

"WebsocketSelector17" prio=10 tid=0x00007f532c0fa000 nid=0x523a runnable [0x00007f5322ad0000]
java.lang.Thread.State: RUNNABLE
at sun.nio.ch.EPollArrayWrapper.epollCtl(Native Method)
at sun.nio.ch.EPollArrayWrapper.updateRegistrations(EPollArrayWrapper.java:246)
- locked <0x000000078b5c06d8> (a java.util.LinkedList)
at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:209)
at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:65)
at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:69)
- locked <0x000000078b5c0600> (a sun.nio.ch.Util$2)
- locked <0x000000078b5c05f0> (a java.util.Collections$UnmodifiableSet)
- locked <0x000000078b5bbea8> (a sun.nio.ch.EPollSelectorImpl)
at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:80)
at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:84)
at org.java_websocket.server.WebSocketServer.run(WebSocketServer.java:298)
at java.lang.Thread.run(Thread.java:662)

[sklegg]

I am seeing a similar issue when many clients are connected. (WebSocketServer.java line 298 waiting for selector.select() to return.)

Doing some research, this method usually churns when multiple threads make register calls or register OP_WRITE when not necessary. I could not find any place in the code where this could be happening.

[Joris-van-der-Wel]

Is it really hanging on select()? (which is okay if there is nothing interesting happening)
Or is it going through the loop in WebSocketServer.java:302 very very often?

If the latter case is true, I am guessing this could be because i.remove() is not called often enough, and the same keys keep getting returned by selector.selectedKeys().

[becast]

I just experienced the same problem but only while using SSL. It is an issue when an SSL session can not fully established or the negotiations break down. Which leaves the Channel in a constant read state. I created a PR.

[marci4]

Hello everyone,
is this still an issue?

Greetings
marci4

[marci4]

Could also be fixed by #488

[marci4]

No feedback for months!

If you still have an issue with that, please open a separate issue again!
Greetings
marci4