Failed to read artifact descriptor for com.alibaba:dubbo:jar:2.6.3

[huadong]

Environment

• Dubbo version: 2.6.3
• Operating System version: Ubuntu 18.04
• Java version: 10.0.1

Step to reproduce this issue

1. create a new project with maven pom.xml depend on dubbo 2.6.3
2. compile the project

Actual Result

Failed to read artifact descriptor for com.alibaba:dubbo:jar:2.6.3: Could not find artifact com.alibaba:dubbo-parent:pom:2.6.3 in central (http://central.maven.org/maven2/)

com.alibaba:dubbo-parent:pom:2.6.3 has not been deployed to http://mvnrepository.com/artifact/com.alibaba/dubbo-parent

but all of others have been deployed to http://mvnrepository.com/artifact/com.alibaba/dubbo

[dslchd]

I have the same question. Is there any official answer?

[diecui1202]

2.6.3 is also on the release way. Lack of dubbo-parent, maybe it's our fault. Pls use 2.6.2 instead.

[dslchd]

I rolled back to version 2.5.10.
Version 2.6.2 also had the shutdown graceful problem in tomcat 8.5.23, and I had to use the following method in my listener:
ProtocolConfig.destroyAll();

[chickenlj]

@huadong The Apache community is still voting for v2.6.3, so it's not available at present. I have updated the release note: https://github.com/apache/incubator-dubbo/releases/tag/untagged-24fcb73b05663c44192a

[huadong]

@chickenlj Thank you!
The tag for v2.6.3 is very helpful, I can try the v2.6.3 from the source code. I think the feature which shutdown graceful is very important, for sometimes we need to use the hot reload feature of tomcat.

[xingqijiuxueyuan]

I have the same question.

[huadong]

@chickenlj I had packaged v2.6.3 from the source code. When shutdown tomcat, it logged this WARNING:

appears to have started a thread named [DubboClientReconnectTimer-thread-2] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
sun.misc.Unsafe.park(Native Method)
java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(ScheduledThreadPoolExecutor.java:1088)
java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(ScheduledThreadPoolExecutor.java:809)
java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
java.lang.Thread.run(Thread.java:748)

So I had to kill the tomcat process instead of executing the shutdown.sh of tomcat.

Should I do any more configurations?

[huadong]

@dslchd According to #1155, the netty blocks shutdown tomcat.
the conflict is :

dubbo v2.6.3 depends on netty v3.2.5,
curator-framework v2.12.0 depends on netty v3.7.0.

tomcat shutdown normally after exclude netty from curator-framework.
Maybe dubbo should explicitly close zookeeper client in the future.

[dslchd]

@huadong Thank you for your answer.
But I'm not using dubbo v2.6.x right now, because I was using v2.6.2 and the graceful shutdown of dubbo was still a problem in tomcat, which is probably why you said it.
I'm now rolling back to v2.5.10 to solve this problem using the following solution
The code is as follows:
ProtocolConfig.destroyAll();
Field field = NettyClient.class.getDeclaredField("channelFactory");
field.setAccessible(true);
ChannelFactory channelFactory = (ChannelFactory) field.get(NettyClient.class);
channelFactory.releaseExternalResources();
field.setAccessible(false);

[ralf0131]

@huadong Please don't mix up the issues, please file a new issue for the elegant shutdown issue.

[huadong]

@ralf0131 Sorry! I filed a new issue #2334 for the elegant shutdown.