[JBoss JIRA] (WFLY-4790) Getting DirectBuffer OOM when sending fragmented binary message to websocket endpoint
by Junshik Jeon (JIRA)
[ https://issues.jboss.org/browse/WFLY-4790?page=com.atlassian.jira.plugin.... ]
Junshik Jeon commented on WFLY-4790:
------------------------------------
I have same issue with WildFly 9.0.2 Final(Undertow 1.2.9). I'm using Wildfly as a websocket endpoint.
I've been tried upgrade Undertow 1.2.12.Final within Wildfly 9.0.2 Final.
But, After that same DirectBuffer OOM error occur.
How can I fix it?
> Getting DirectBuffer OOM when sending fragmented binary message to websocket endpoint
> -------------------------------------------------------------------------------------
>
> Key: WFLY-4790
> URL: https://issues.jboss.org/browse/WFLY-4790
> Project: WildFly
> Issue Type: Bug
> Components: Web (Undertow)
> Affects Versions: 10.0.0.Alpha3
> Reporter: Radim Hatlapatka
> Assignee: Stuart Douglas
> Priority: Critical
> Fix For: 10.0.0.Alpha6
>
>
> When sending fragmented binary message (message with message payload of length 4 * 2**20 (4M). Sent out in fragments of 64). The server throws {{java.lang.OutOfMemoryError: Direct buffer memory}} [1]
> The memory for direct buffer by default depends on the size set by -Xmx, which is in EAP 7.0.0.DR4 by default set to -Xmx512m. Increasing it just increases the time before the limit is hit (it is enough to send those messages multiple times to hit the limit again).
> I believe the issue is similar to the one for EAP 6.4: [https://bugzilla.redhat.com/show_bug.cgi?id=1223708]
> [1]
> {noformat}
> 15:10:55,463 ERROR [org.xnio.listener] (default I/O-1) XNIO001007: A channel event listener threw an exception: java.lang.OutOfMemoryError: Direct buffer memory
> at java.nio.Bits.reserveMemory(Bits.java:658)
> at java.nio.DirectByteBuffer.<init>(DirectByteBuffer.java:123)
> at java.nio.ByteBuffer.allocateDirect(ByteBuffer.java:311)
> at org.xnio.BufferAllocator$2.allocate(BufferAllocator.java:57)
> at org.xnio.BufferAllocator$2.allocate(BufferAllocator.java:55)
> at org.xnio.ByteBufferSlicePool.allocate(ByteBufferSlicePool.java:143)
> at io.undertow.websockets.core.BufferedBinaryMessage$1.handleEvent(BufferedBinaryMessage.java:106)
> at io.undertow.websockets.core.BufferedBinaryMessage$1.handleEvent(BufferedBinaryMessage.java:97)
> at org.xnio.ChannelListeners.invokeChannelListener(ChannelListeners.java:92)
> at io.undertow.server.protocol.framed.AbstractFramedStreamSourceChannel$1.run(AbstractFramedStreamSourceChannel.java:264)
> at org.xnio.nio.WorkerThread.safeRun(WorkerThread.java:560)
> at org.xnio.nio.WorkerThread.run(WorkerThread.java:462)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
8 years, 5 months
[JBoss JIRA] (WFLY-5921) Locale cache set to simple cache automatically breaks map reduce
by Paul Ferraro (JIRA)
[ https://issues.jboss.org/browse/WFLY-5921?page=com.atlassian.jira.plugin.... ]
Paul Ferraro closed WFLY-5921.
------------------------------
Resolution: Rejected
Infinispan map-reduce API is scheduled for deprecation (in 8.x) in favor of distributed streams. Migrating to this API should be fairly trivial. See:
https://github.com/infinispan/infinispan/wiki/Distributed-Stream-Support
> Locale cache set to simple cache automatically breaks map reduce
> ----------------------------------------------------------------
>
> Key: WFLY-5921
> URL: https://issues.jboss.org/browse/WFLY-5921
> Project: WildFly
> Issue Type: Bug
> Components: Clustering
> Affects Versions: 10.0.0.CR5
> Reporter: Stian Thorgersen
> Assignee: Paul Ferraro
> Priority: Critical
>
> We use the Infinispan subsystem to create/configure caches for Keycloak. In standalone mode Keycloak uses locale-caches, while in clustered mode we use a combination of invalidation caches and distributed caches. In both clustered and non-clustered mode we use map-reduce tasks to delete elements from the cache.
> This has worked just fine until the recent change in CR5 where locale-caches are now set to simple-caches (WFLY-5327). As the caches are now automatically set to simple-caches map-reduce is no longer available. There's also no way to prevent this.
> IMO the changes from WFLY-5327 should be reverted and instead a new simple-cache element should be added, or a simple-cache=true attribute added to the locale-cache element.
> As it stands this change prevents us from upgrading to CR5.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
8 years, 5 months
[JBoss JIRA] (DROOLS-847) WebLogic unable to deploy Execution server
by Lukáš Petrovický (JIRA)
[ https://issues.jboss.org/browse/DROOLS-847?page=com.atlassian.jira.plugin... ]
Lukáš Petrovický updated DROOLS-847:
------------------------------------
Labels: reported-by-qe (was: )
> WebLogic unable to deploy Execution server
> ------------------------------------------
>
> Key: DROOLS-847
> URL: https://issues.jboss.org/browse/DROOLS-847
> Project: Drools
> Issue Type: Sub-task
> Components: kie server
> Affects Versions: 6.3.0.Beta2
> Environment: WebLogic
> Reporter: Karel Suta
> Assignee: Karel Suta
> Labels: reported-by-qe
> Fix For: 6.3.0.Final
>
>
> Unified Execution server WAR doesn't contain dependency on Hibernate needed by jBPM part. This is ok for Wildfly which has its own Hibernate module included, but WebLogic and possibly Websphere doesn't contain Hibernate causing ClassNotFoundException when deploying exec server.
> One of possible solution would be to create separate war for WebLogic and Websphere, as is done for Business Central, containing container specific resources.
> Tried on ee6 WAR, same as was used in previous release.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
8 years, 5 months
[JBoss JIRA] (DROOLS-948) testEventWithShortExpiration() in CepEspTest and CepJavaTypeTest fails
by Lukáš Petrovický (JIRA)
[ https://issues.jboss.org/browse/DROOLS-948?page=com.atlassian.jira.plugin... ]
Lukáš Petrovický updated DROOLS-948:
------------------------------------
Labels: reported-by-qe (was: )
> testEventWithShortExpiration() in CepEspTest and CepJavaTypeTest fails
> ----------------------------------------------------------------------
>
> Key: DROOLS-948
> URL: https://issues.jboss.org/browse/DROOLS-948
> Project: Drools
> Issue Type: Bug
> Reporter: Tibor Zimányi
> Assignee: Mark Proctor
> Priority: Minor
> Labels: reported-by-qe
>
> Last assert in testEventWithShortExpiration() sometimes fails because sometimes the event don't have time to expire (although there is 1ms expiration defined) after fireAllRules is called in the test before the assert. There should be some sleep time added before the assert.
> The fail is nondeterministic and this nondeterminism is because 1ms expire time can be longer because of jvm scheduler. Issue can be reproduced by running the test in a for loop with e.g. 1000 iterations.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
8 years, 5 months