atom feed3 messages in org.apache.hadoop.zookeeper-userRe: Thread leak problem
FromSent OnAttachments
Justyanzi LinMar 4, 2012 1:58 am 
Patrick HuntMar 6, 2012 9:08 am 
Ted DunningMar 6, 2012 9:11 am 
Subject:Re: Thread leak problem
From:Ted Dunning (ted.@gmail.com)
Date:Mar 6, 2012 9:11:07 am
List:org.apache.hadoop.zookeeper-user

If you were able to connect to a 1 out of 3 servers, then you have a very serious problem since that isn't enough machines to form a quorum.

I suspect you have a configuration error.

On Tue, Mar 6, 2012 at 9:08 AM, Patrick Hunt <phu@apache.org> wrote:

This is windows I take it? What version of zk was used, what compiler was used, etc... those sorts of details would be helpful to nail this down. Windows (visual studio) support is relatively new, perhaps the threads are not being closed down properly for some reason. Do the logs give any indication? (try turning up the logging level on the client to get more detail).

Patrick

Patrick

On Sun, Mar 4, 2012 at 1:59 AM, Justyanzi Lin <just@gmail.com> wrote:

I modify cli.c let it only run a for-loop.

the loop as below

while(1) { zh = zookeeper_init(hostPort, watcher, 30000, &myid, 0, 0); Sleep(1000); zookeeper_close(zh); }

And I found every time execute zookeeper_close(zh), will leave 2 thread handles. I support to thread handle will the same.

It seems not correct.

I use WinDbg tool to watch the process and found first stop 156 Handles Type Count None 2 Event 15 File 2 Directory 3 Mutant 1 WindowStation 2 Semaphore 2 Key 6 Thread 121 Desktop 1 IoCompletion 1

second stop 193 Handles Type Count None 2 Event 17 File 2 Directory 3 Mutant 1 WindowStation 2 Semaphore 2 Key 6 Thread 156 Desktop 1 IoCompletion 1

Is there something I using wrong?

Thanks