[
https://issues.apache.org/jira/browse/NUTCH-2936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17554531#comment-17554531
]
Sebastian Nagel commented on NUTCH-2936:
----------------------------------------
After debugging this: the call by the Hadoop MR Job to initialize the
KeyGenerator leads twice recursively into Nutch's URLStreamHandlerFactory -
first for the "http" protocol to create a [NULL_URL
(http://null.oracle.com/)|https://github.com/openjdk/jdk/blob/0530f4e517be5d5b3ff10be8a0764e564f068c06/src/java.base/share/classes/javax/crypto/JceSecurity.java.template#L246],
second for the "jar" to load the protocol-okhttp.jar. See the debug log output
and the stack trace (some lines stripped):
{noformat}
2022-06-14 16:56:59,176 DEBUG plugin.URLStreamHandlerFactory: Registered
URLStreamHandlerFactory with the JVM.
2022-06-14 16:56:59,994 DEBUG plugin.URLStreamHandlerFactory: Creating
URLStreamHandler for protocol: http
2022-06-14 16:56:59,994 DEBUG plugin.PluginRepository: Creating
URLStreamHandler for protocol: http
2022-06-14 16:56:59,995 DEBUG plugin.PluginRepository: Suitable protocolName
attribute located: http
2022-06-14 16:57:00,007 DEBUG plugin.URLStreamHandlerFactory: Creating
URLStreamHandler for protocol: jar
2022-06-14 16:57:00,007 DEBUG plugin.PluginRepository: Creating
URLStreamHandler for protocol: jar
2022-06-14 16:57:00,008 DEBUG plugin.PluginRepository: No suitable protocol
extensions registered for protocol: jar
2022-06-14 16:57:00,320 DEBUG plugin.PluginRepository: Located extension
instance class: org.apache.nutch.protocol.okhttp.OkHttp
2022-06-14 16:57:00,320 DEBUG plugin.PluginRepository: Suitable protocol
extension found that did not declare a handler
{noformat}
{noformat}
at
org.apache.nutch.plugin.PluginRepository.createURLStreamHandler(PluginRepository.java:583)
at
org.apache.nutch.plugin.URLStreamHandlerFactory.createURLStreamHandler(URLStreamHandlerFactory.java:95)
at java.base/java.net.URL.getURLStreamHandler(URL.java:1432)
at java.base/java.net.URL.<init>(URL.java:451)
at
java.base/jdk.internal.loader.URLClassPath$JarLoader.<init>(URLClassPath.java:720)
at
java.base/jdk.internal.loader.URLClassPath$3.run(URLClassPath.java:494)
at
java.base/jdk.internal.loader.URLClassPath$3.run(URLClassPath.java:477)
at java.base/java.security.AccessController.doPrivileged(Native Method)
at
java.base/jdk.internal.loader.URLClassPath.getLoader(URLClassPath.java:476)
at
java.base/jdk.internal.loader.URLClassPath.getLoader(URLClassPath.java:445)
at
java.base/jdk.internal.loader.URLClassPath.getResource(URLClassPath.java:314)
at java.base/java.net.URLClassLoader$1.run(URLClassLoader.java:455)
at java.base/java.net.URLClassLoader$1.run(URLClassLoader.java:452)
at java.base/java.security.AccessController.doPrivileged(Native Method)
at java.base/java.net.URLClassLoader.findClass(URLClassLoader.java:451)
at
org.apache.nutch.plugin.PluginClassLoader.loadClass(PluginClassLoader.java:71)
at java.base/java.lang.ClassLoader.loadClass(ClassLoader.java:522)
at
org.apache.nutch.plugin.PluginRepository.getCachedClass(PluginRepository.java:349)
at
org.apache.nutch.plugin.Extension.getExtensionInstance(Extension.java:165)
at
org.apache.nutch.plugin.PluginRepository.createURLStreamHandler(PluginRepository.java:601)
at
org.apache.nutch.plugin.URLStreamHandlerFactory.createURLStreamHandler(URLStreamHandlerFactory.java:95)
at java.base/java.net.URL.getURLStreamHandler(URL.java:1432)
at java.base/java.net.URL.<init>(URL.java:651)
at java.base/java.net.URL.<init>(URL.java:541)
at java.base/java.net.URL.<init>(URL.java:488)
at java.base/javax.crypto.JceSecurity.<clinit>(JceSecurity.java:239)
at java.base/javax.crypto.KeyGenerator.nextSpi(KeyGenerator.java:363)
at java.base/javax.crypto.KeyGenerator.<init>(KeyGenerator.java:176)
at
java.base/javax.crypto.KeyGenerator.getInstance(KeyGenerator.java:244)
at
org.apache.hadoop.mapreduce.JobSubmitter.submitJobInternal(JobSubmitter.java:179)
...
at org.apache.hadoop.mapreduce.Job.submit(Job.java:1568)
at org.apache.hadoop.mapreduce.Job.waitForCompletion(Job.java:1589)
at org.apache.nutch.crawl.Injector.inject(Injector.java:436)
at org.apache.nutch.crawl.Injector.run(Injector.java:569)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:81)
at org.apache.nutch.crawl.Injector.main(Injector.java:533)
...
{noformat}
I do not understand why the initialization of the KeyGenerator fails only in
this combination (distributed mode and using protocol-okhttp). Nevertheless, we
should never delegate standard URLStreamHandlers implemented by the JVM to
handlers requiring the Nutch plugin system with its complexity and the
plugin-specific class loaders. This may break stuff unexpectedly, esp. if
stream handlers are used to open connections. See also HADOOP-14598 for a
similar issue.
> Early registration of URL stream handlers provided by plugins may fail Hadoop
> jobs running in distributed mode if protocol-okhttp is used
> -----------------------------------------------------------------------------------------------------------------------------------------
>
> Key: NUTCH-2936
> URL: https://issues.apache.org/jira/browse/NUTCH-2936
> Project: Nutch
> Issue Type: Bug
> Components: plugin, protocol
> Affects Versions: 1.19
> Reporter: Sebastian Nagel
> Assignee: Lewis John McGibbney
> Priority: Blocker
> Fix For: 1.19
>
>
> After merging NUTCH-2429 I've observed that Nutch jobs running in distributed
> mode may fail early with the following dubious error:
> {noformat}
> 2022-01-14 13:11:45,751 ERROR crawl.DedupRedirectsJob: DeduplicationJob:
> java.io.IOException: Error generating shuffle secret key
> at
> org.apache.hadoop.mapreduce.JobSubmitter.submitJobInternal(JobSubmitter.java:182)
> at org.apache.hadoop.mapreduce.Job$11.run(Job.java:1565)
> at org.apache.hadoop.mapreduce.Job$11.run(Job.java:1562)
> at java.base/java.security.AccessController.doPrivileged(Native
> Method)
> at java.base/javax.security.auth.Subject.doAs(Subject.java:423)
> at
> org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1762)
> at org.apache.hadoop.mapreduce.Job.submit(Job.java:1562)
> at org.apache.hadoop.mapreduce.Job.waitForCompletion(Job.java:1583)
> at
> org.apache.nutch.crawl.DedupRedirectsJob.run(DedupRedirectsJob.java:301)
> at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76)
> at
> org.apache.nutch.crawl.DedupRedirectsJob.main(DedupRedirectsJob.java:379)
> at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
> at
> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.base/java.lang.reflect.Method.invoke(Method.java:566)
> at org.apache.hadoop.util.RunJar.run(RunJar.java:323)
> at org.apache.hadoop.util.RunJar.main(RunJar.java:236)
> Caused by: java.security.NoSuchAlgorithmException: HmacSHA1 KeyGenerator not
> available
> at java.base/javax.crypto.KeyGenerator.<init>(KeyGenerator.java:177)
> at
> java.base/javax.crypto.KeyGenerator.getInstance(KeyGenerator.java:244)
> at
> org.apache.hadoop.mapreduce.JobSubmitter.submitJobInternal(JobSubmitter.java:179)
> ... 16 more
> {noformat}
> After removing the early registration of URL stream handlers (see NUTCH-2429)
> in NutchJob and NutchTool, the job starts without errors.
> Notes:
> - the job this error was observed a [custom de-duplication
> job|https://github.com/commoncrawl/nutch/blob/cc/src/java/org/apache/nutch/crawl/DedupRedirectsJob.java]
> to flag redirects pointing to the same target URL. But I'll try to reproduce
> it with a standard Nutch job and in pseudo-distributed mode.
> - should also verify whether registering URL stream handlers works at all in
> distributed mode. Tasks are launched differently, not as NutchJob or
> NutchTool.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
