Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Accumulo >> mail # user >> PermGen leak in JBoss


+
Michael Giordano 2013-04-24, 13:48
+
Heath Abelson 2013-04-24, 15:21
+
Keith Turner 2013-04-24, 16:22
+
Michael Giordano 2013-05-03, 15:43
+
Keith Turner 2013-05-03, 15:54
+
Keith Turner 2013-04-24, 15:44
+
Eric Newton 2013-04-24, 15:47
+
Heath Abelson 2013-04-24, 16:08
+
Keith Turner 2013-04-24, 16:34
+
Benson Margulies 2013-04-24, 16:58
+
Heath Abelson 2013-04-24, 16:58
+
Josh Elser 2013-04-24, 14:12
Copy link to this message
-
Re: PermGen leak in JBoss
On Wed, Apr 24, 2013 at 10:12 AM, Josh Elser <[EMAIL PROTECTED]> wrote:

> Giving a quick glance at that code again, the only thing I see that would
> be active is the ZooKeeperInstance (ZooCache) which will try to stay
> connected/updated to ZooKeeper. However, that should also disconnect/expire
> itself.
>
> What's the time frame in which you're redeploying your webapp (within a
> few seconds, minutes)? Also, threadname/stacktrace for said 5 threads would
> be helpful/enlightening.
Yeah those stack traces would be nice if you have them.  I suspect one
culprit may come from o.a.accumulo.core.client.impl.ThriftTransportPool,
but I am curious about the stack traces.  I am going to try firing up a
client and doing what you did in 1.4 and looking at the stacks.
>
>
> On 4/24/13 9:48 AM, Michael Giordano wrote:
>
>> Under version 1.3.7 we are using the following code to initialize a
>> cloudbase connection during initialization of our web app:
>>
>>                         ZooKeeperInstance instance = new
>> ZooKeeperInstance(**instanceName, zooKeepers);
>>                         connector = instance.getConnector(userId,
>> password.getBytes());
>>
>> The problem is that under the hood, this call creates several (5) threads
>> that are not cleaned up when the app is undeployed in JBoss. This is
>> occurring without performing any scans or interacting with cloudbase in any
>> other way. After relatively few redeploys of the app, the PermGen Space is
>> OOM.
>>
>> I can't find any reference in the cloudbase API akin to a close() method
>> for the Connector object. This is a classloader leak effecting any webapp
>> that is accessing cloudbase directly. The result of this leak is not simply
>> orphaned threads, but thousands of classes not gc'd because the classloader
>> itself can't be gc'd. This is what is filling up PermGen.
>>
>> Has anyone discovered this particular issue? Has anyone discovered a
>> solution?
>>
>> Thanks,
>> Mike G.
>>
>> NetCentric Technology, Inc.
>> 3349 Route 138, Building A
>> Wall, NJ  07719
>> Phone: 732-544-0888
>>
>>
>>
>