[HADOOP-15865] ConcurrentModificationException in Configuration.overlay() method

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 3.3.0, 3.1.4, 3.2.2
• Assignee: Oleksandr Shevchenko
• Reporter: Oleksandr Shevchenko
• Created: 19/Oct/18 07:50
• Updated: 01/Oct/19 00:33
• Resolved: 02/Aug/19 02:57

Description

Configuration.overlay() is not thread-safe and can be the cause of ConcurrentModificationException since we use iteration over Properties object.

private void overlay(Properties to, Properties from) {
 for (Entry<Object, Object> entry: from.entrySet()) {
 to.put(entry.getKey(), entry.getValue());
 }
 }

Properties class is thread-safe but iterator is not. We should manually synchronize on the returned set of entries which we use for iteration.

We faced with ResourceManger fails during recovery caused by ConcurrentModificationException:

2018-10-12 08:00:56,968 INFO org.apache.hadoop.service.AbstractService: Service ResourceManager failed in state STARTED; cause: java.util.ConcurrentModificationException
java.util.ConcurrentModificationException
 at java.util.Hashtable$Enumerator.next(Hashtable.java:1383)
 at org.apache.hadoop.conf.Configuration.overlay(Configuration.java:2801)
 at org.apache.hadoop.conf.Configuration.loadResource(Configuration.java:2696)
 at org.apache.hadoop.conf.Configuration.loadResources(Configuration.java:2632)
 at org.apache.hadoop.conf.Configuration.getProps(Configuration.java:2528)
 at org.apache.hadoop.conf.Configuration.get(Configuration.java:1062)
 at org.apache.hadoop.conf.Configuration.getStringCollection(Configuration.java:1914)
 at org.apache.hadoop.security.alias.CredentialProviderFactory.getProviders(CredentialProviderFactory.java:53)
 at org.apache.hadoop.conf.Configuration.getPasswordFromCredentialProviders(Configuration.java:2043)
 at org.apache.hadoop.conf.Configuration.getPassword(Configuration.java:2023)
 at org.apache.hadoop.yarn.webapp.util.WebAppUtils.getPassword(WebAppUtils.java:452)
 at org.apache.hadoop.yarn.webapp.util.WebAppUtils.loadSslConfiguration(WebAppUtils.java:428)
 at org.apache.hadoop.yarn.webapp.WebApps$Builder.start(WebApps.java:293)
 at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.startWepApp(ResourceManager.java:1017)
 at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.serviceStart(ResourceManager.java:1117)
 at org.apache.hadoop.service.AbstractService.start(AbstractService.java:193)
 at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.main(ResourceManager.java:1251)
2018-10-12 08:00:56,968 INFO org.apache.hadoop.yarn.server.resourcemanager.security.RMDelegationTokenSecretManager: removing RMDelegation token with sequence number: 3489914
2018-10-12 08:00:56,968 INFO org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore: Removing RMDelegationToken and SequenceNumber
2018-10-12 08:00:56,968 INFO org.apache.hadoop.yarn.server.resourcemanager.recovery.FileSystemRMStateStore: Removing RMDelegationToken_3489914
2018-10-12 08:00:56,969 INFO org.apache.hadoop.ipc.Server: Stopping server on 8032