|
| Re: [CDO] Multi master (clone) [message #923816 is a reply to message #923803] |
Wed, 26 September 2012 08:08   |
|
Am 26.09.2012 09:57, schrieb Ricky de Klerck:
> I was told to ask this question on the emf forum so here it is.. I managed to create a CDO master and clone following
> the examples in org.eclipse.emf.cdo.examples. (I'm trying to achieve the architecture described like
> http://www.eclipse.org/forums/index.php/m/880089/)
>
> I still have one problem now.. I've set up 2 clients, one connected to the master and one to the clone. When I save an
> object in the clone, everything is synchronized well, but the other way around (save object in master) doesn't work.
> The changes aren't pushed (or pulled) from master to clone. Anybody had this problem before?
The IRepositorySynchronizer of the CloneRepository registers with the master repository, which then will push all
changes (see PassiveUpdateMode.ADDITIONS) to the registered clone. If commits to the master are not pushed to the clone
then you're either doing something wrong or there's a bug in CDO's replication mechansim. The latter is less likely (no
complaints so far and tests are all passing) but it's impossible for me to judge without more information. Maybe you can
submit a bugzilla and attach a small, *executable* example?
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
| Re: [CDO] Multi master (clone) [message #923993 is a reply to message #923816] |
Wed, 26 September 2012 11:39 |
 Ricky de Klerck
Messages: 295
Registered: January 2011 |
Senior Member |
|
|
Thanks for your answer. Creating an executable example is not that easy because I now have 2 seperate clients, a clone and a master.. So before I do that, some code might do the trick. First of all, the 'PassiveUpdateMode.ADDITIONS' you mentioned is part of the client, right? The session is created and a setPassiveUpdateMode() is done?
The code of my clients is identical, only the repository is different (clone or master):
System.out.println("Client starting...");
IManagedContainer container = OfflineExampleUtil.createContainer();
IConnector connector = Net4jUtil.getConnector(container, AbstractOfflineExampleServer.TRANSPORT_TYPE, "localhost:"
+ 2036);
CDONet4jSessionConfiguration configuration = CDONet4jUtil.createNet4jSessionConfiguration();
configuration.setConnector(connector);
configuration.setRepositoryName("cdo");
CDONet4jSession session = configuration.openNet4jSession();
session.options().setPassiveUpdateEnabled(true);
session.options().setPassiveUpdateMode(PassiveUpdateMode.ADDITIONS);
// session.getPackageRegistry().putEPackage(VlaamsparlementPackage.eINSTANCE);
CDORepositoryInfo repositoryInfo = session.getRepositoryInfo();
System.out.println("Connected to " + repositoryInfo.getName());
return session;
This is the code to create the master:
CDOServerUtil.createRepository(name, store, props);
And this is (some of) the code to create the clone:
...
CDOServerUtil.createOfflineClone(name, store, props, synchronizer);
...
protected IRepositorySynchronizer createRepositorySynchronizer(String connectorDescription, String repositoryName)
{
CDOSessionConfigurationFactory factory = createSessionConfigurationFactory(connectorDescription, repositoryName);
IRepositorySynchronizer synchronizer = CDOServerUtil.createRepositorySynchronizer(factory);
synchronizer.setRetryInterval(2);
synchronizer.setMaxRecommits(10);
synchronizer.setRecommitInterval(2);
synchronizer.setRawReplication(true);
return synchronizer;
}
I make use of the AbstractOfflineExampleServer which is used in org.eclipse.emf.cdo.examples.
Does this help to solve the problem in any way?
[Updated on: Wed, 26 September 2012 12:08] Report message to a moderator |
|
|
| Re: [CDO] Multi master (clone) [message #924802 is a reply to message #923993] |
Thu, 27 September 2012 06:16 |
|
Am 26.09.2012 13:39, schrieb Ricky de Klerck:
> Thanks for your answer. To create an executable example is not that easy because I now have 2 seperate clients, a
> clone and a master.. So before I do that, some code might do the trick. First of all, the
> 'PassiveUpdateMode.ADDITIONS' you mentioned is part of the client, right? The session is created and a
> setPassiveUpdateMode() is done?
No, in a "normal" client you don't need PassiveUpdateMode.ADDITIONS and probably shouldn't use it as it leads to
substantially more network traffic. I just mentioned it because the "client" in the RepositorySynchronizer uses this
mode to receive all changes and additions from the master:
CDOSessionConfiguration masterConfiguration = remoteSessionConfigurationFactory.createSessionConfiguration();
masterConfiguration.setPassiveUpdateMode(PassiveUpdateMode.ADDITIONS);
masterConfiguration.setLockNotificationMode(LockNotificationMode.ALWAYS);
remoteSession = (InternalCDOSession)masterConfiguration.openSession();
ensureNOOPRevisionCache();
> The code of my clients is identical, only the repository is different (clone or master):
That should be okay.
> System.out.println("Client starting...");
> IManagedContainer container = OfflineExampleUtil.createContainer();
> IConnector connector = Net4jUtil.getConnector(container, AbstractOfflineExampleServer.TRANSPORT_TYPE, "localhost:"
> + 2036);
The port is correct?
> CDONet4jSessionConfiguration configuration = CDONet4jUtil.createNet4jSessionConfiguration();
> configuration.setConnector(connector);
> configuration.setRepositoryName("cdo");
>
> CDONet4jSession session = configuration.openNet4jSession();
> session.options().setPassiveUpdateEnabled(true);
> session.options().setPassiveUpdateMode(PassiveUpdateMode.ADDITIONS);
It's a little bit faster to set these already on the configuration used to open the session.
> // session.getPackageRegistry().putEPackage(VlaamsparlementPackage.eINSTANCE);
putEPackage() is indeed not needed anymore for generated and deployed packages with post 3.0 versions of CDO.
> CDORepositoryInfo repositoryInfo = session.getRepositoryInfo();
> System.out.println("Connected to " + repositoryInfo.getName());
>
> return session;
>
> This is the code to create the master:
> CDOServerUtil.createRepository(name, store, props);
>
> And this is (some of) the code to create the clone:
>
> ..
> CDOServerUtil.createOfflineClone(name, store, props, synchronizer);
> ..
>
> protected IRepositorySynchronizer createRepositorySynchronizer(String connectorDescription, String repositoryName)
> {
> CDOSessionConfigurationFactory factory = createSessionConfigurationFactory(connectorDescription, repositoryName);
>
> IRepositorySynchronizer synchronizer = CDOServerUtil.createRepositorySynchronizer(factory);
> synchronizer.setRetryInterval(2);
> synchronizer.setMaxRecommits(10);
> synchronizer.setRecommitInterval(2);
> synchronizer.setRawReplication(true);
> return synchronizer;
> }
That looks okay to me.
> I make use of the AbstractOfflineExampleServer which is used in org.eclipse.emf.cdo.examples.
>
> Does this help to solve the problem in any way?
I don't think so ;-(
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
| Re: [CDO] Multi master (clone) [message #925095 is a reply to message #924980] |
Thu, 27 September 2012 11:15 |
|
Am 27.09.2012 11:30, schrieb Ricky de Klerck:
> The ports are correct, otherwise it shouldn't synchronize the other way i suppose.
>
> Is there a certain way a transaction should be started/commited? I now have a method to create test data like this (i
> deleted a whole bunch btw):
>
>
> public static void createDummyData(){
> try {
> CDOTransaction transaction = cdoSession.openTransaction();
>
> CDOResource parlementsResource = transaction.getOrCreateResource("/parlements");
> CDOResource personsResource = transaction.getOrCreateResource("/persons");
>
> if(parlementsResource.getContents().size() == 0){
> Parlement parlement = VlaamsparlementFactory.eINSTANCE.createParlement();
> //SPA
> Party spa = VlaamsparlementFactory.eINSTANCE.createParty();
> spa.setName("SPA");
> spa.setAddress("test 105/37\n" +
> "1000 test");
> parlement.getParties().add(spa);
> Person vdl = VlaamsparlementFactory.eINSTANCE.createPerson();
> vdl.setFirstName("Jan");
> vdl.setLastName("Achternaam");
> vdl.setGender(Genders.MALE);
> vdl.setDateOfBirth(new Date());
> spa.getPersons().add(vdl);
>
> Person li = VlaamsparlementFactory.eINSTANCE.createPerson();
> li.setFirstName("Lotte");
> li.setLastName("Achternaam");
> li.setGender(Genders.FEMALE);
> li.setDateOfBirth(new Date());
> spa.getPersons().add(li);
>
> spa.setLeader(vdl);
>
> parlementsResource.getContents().add(parlement);
>
> personsResource.getContents().add(vdl);
> personsResource.getContents().add(li);
>
>
> transaction.commit();
> }
> transaction.close();
> } catch (CommitException e) {
> e.printStackTrace();
> }
> }
>
>
> Is this the correct way?
Looks okay to me. As you don't use explicit write locks for your modifications I suggest to synchronize the entire try
block on the used transaction. Another enhancement would be to repeat that block until there's no CommitException anymore.
> After the commit (client connected to master), it should be synchronized with the clone database, right?
Yes, that should happen. And it does in our offline test suite, which I just double checked.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
| Re: [CDO] Multi master (clone) [message #925229 is a reply to message #925095] |
Thu, 27 September 2012 14:10 |
Ricky de Klerck
Messages: 295
Registered: January 2011 |
Senior Member |
|
|
Ok, i did a little test with the testsuite from org.eclipse.emf.cdo.examples (grabbed the source from http://git.eclipse.org/c/cdo/cdo.git/tree/plugins/org.eclipse.emf.cdo.examples/src/org/eclipse/emf/cdo/examples/server/offline). I tested it with mysql databases so this is the code I changed:
com.mysql.jdbc.jdbc2.optional.MysqlDataSource ds = new com.mysql.jdbc.jdbc2.optional.MysqlDataSource();
ds.setUrl("jdbc:mysql://localhost/" + name);
ds.setUser("cdo");
ds.setPassword("cdo");
IMappingStrategy mappingStrategy = CDODBUtil.createHorizontalMappingStrategy(true, true);
IDBAdapter dbAdapter = new MYSQLAdapter();
IDBConnectionProvider dbConnectionProvider = DBUtil.createConnectionProvider(ds);
return CDODBUtil.createStore(mappingStrategy, dbAdapter, dbConnectionProvider);
//JdbcDataSource dataSource = new JdbcDataSource();
//dataSource.setURL("jdbc:h2:database/" + name);
//
//IMappingStrategy mappingStrategy = CDODBUtil.createHorizontalMappingStrategy(true, true);
//IDBAdapter dbAdapter = new H2Adapter();
//IDBConnectionProvider dbConnectionProvider = DBUtil.createConnectionProvider(dataSource);
//return CDODBUtil.createStore(mappingStrategy, dbAdapter, dbConnectionProvider);
Test 1:
- removed all tables from both databases (clone and master)
- started master (all tables are created)
- started clone (all tables are created)
- started client (CONNECTED TO CLONE)
- added object via client
- object is added in both databases
- Result: OK
Test 2:
- removed all tables from both databases (clone and master)
- started master (all tables are created)
- started clone (all tables are created)
- started client (CONNECTED TO MASTER)
- added object via client
- object is only added in master database
- Result: NOT OK
[Updated on: Thu, 27 September 2012 14:11] Report message to a moderator |
|
|
|
|
| Re: [CDO] Multi master (clone) [message #926062 is a reply to message #926050] |
Fri, 28 September 2012 09:11 |
|
Am 28.09.2012 10:59, schrieb Ricky de Klerck:
> I installed the "CDO Model Repository Tests" feature and configured it to work with my mysql database, but the test i
> described earlier still doesn't work. In the ExampleClient I started a CDO session to my MASTER and opened a
> transaction. In this transaction, I create a resource and add a Category object.
> After the save the object is stored in the MASTER database but not in the CLONE. The other way around (client
> connected to clone) works fine.. Is this the exact scenario you tested? (except mysql database)
Yes: org.eclipse.emf.cdo.tests.offline.OfflineTest.testMasterCommits_ArrivalInClone()
In the 4.2 branch (Git "master") I experience a problem with the root resource ID which I need to analyze/fix. But I
doubt that it's related to your problem. What CDO version are you using?
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
| Re: [CDO] Multi master (clone) [message #926317 is a reply to message #925229] |
Fri, 28 September 2012 14:23 |
|
Am 27.09.2012 16:10, schrieb Ricky de Klerck:
> Ok, i did a little test with the testsuite from org.eclipse.emf.cdo.examples (grabbed the source from http://git.eclipse.org/c/cdo/cdo.git/tree/plugins/org.eclipse.emf.cdo.examples/src/org/eclipse/emf/cdo/examples/server/offline). I tested it with mysql databases so this is the code I changed:
>
>
> com.mysql.jdbc.jdbc2.optional.MysqlDataSource ds = new com.mysql.jdbc.jdbc2.optional.MysqlDataSource();
> ds.setUrl("jdbc:mysql://localhost/" + name);
> ds.setUser("cdo");
> ds.setPassword("cdo");
>
> IMappingStrategy mappingStrategy = CDODBUtil.createHorizontalMappingStrategy(true, true);
> IDBAdapter dbAdapter = new MYSQLAdapter();
> IDBConnectionProvider dbConnectionProvider = DBUtil.createConnectionProvider(ds);
> return CDODBUtil.createStore(mappingStrategy, dbAdapter, dbConnectionProvider);
>
> //JdbcDataSource dataSource = new JdbcDataSource();
> //dataSource.setURL("jdbc:h2:database/" + name);
> //
> //IMappingStrategy mappingStrategy = CDODBUtil.createHorizontalMappingStrategy(true, true);
> //IDBAdapter dbAdapter = new H2Adapter();
> //IDBConnectionProvider dbConnectionProvider = DBUtil.createConnectionProvider(dataSource);
> //return CDODBUtil.createStore(mappingStrategy, dbAdapter, dbConnectionProvider);
>
>
>
>
> Test 1:
>
> - removed all tables from both databases (clone and master)
> - started master (all tables are created)
> - started clone (all tables are created)
> - started client (CONNECTED TO CLONE)
> - added object via client
> - object is added in both databases
> - Result: OK
>
> Test 2:
>
> - removed all tables from both databases (clone and master)
> - started master (all tables are created)
> - started clone (all tables are created)
> - started client (CONNECTED TO MASTER)
> - added object via client
> - object is only added in master database
> - Result: NOT OK
I've tried both things and many others in both CDO branches 4.2 integartion and 4.1 maintenance. It works perfectly with
H2. Have you tried it with H2?
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
|
|
|
| Re: [CDO] Multi master (clone) [message #929298 is a reply to message #929290] |
Mon, 01 October 2012 10:58 |
|
Am 01.10.2012 12:52, schrieb Ricky de Klerck:
> Can you please describe the steps you did to do the test?
Test 1:
- removed all tables from both databases (clone and master)
- started master (all tables are created)
- started clone (all tables are created)
- started client (CONNECTED TO CLONE)
- added object via client
- object is added in both databases
- Result: OK
Test 2:
- removed all tables from both databases (clone and master)
- started master (all tables are created)
- started clone (all tables are created)
- started client (CONNECTED TO MASTER)
- added object via client
- object is added in both databases
- Result: OK
> On the moment I start the test, only CDO tables are present in the master/clone.. When I add an object in the master,
> the tables from the EMF package are created (in the master). Is this also the case with your test?
I didn't check but without them the test would not pass, right?
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
|
|
| Re: [CDO] Multi master (clone) [message #929540 is a reply to message #929430] |
Mon, 01 October 2012 14:50 |
|
Am 01.10.2012 15:04, schrieb Ricky de Klerck:
> Finally! That did the trick. I didn't have this line in my clone configuration:
>
>
> Map<String, String> props = new HashMap<String, String>();
> props.put(IMappingStrategy.PROP_QUALIFIED_NAMES, "true");
> mappingStrategy.setProperties(props);
Good that it works now.
> But how can it be that this line isn't in the examples? I dubble checked but it isn't there (file - new - Clone server
> example). So the example you ran didn't have this line of code, but still worked?
Hmm, I used these classes directly from my development IDE:
/org.eclipse.emf.cdo.examples/src/org/eclipse/emf/cdo/examples/server/offline/OfflineExampleClone.java
/org.eclipse.emf.cdo.examples/src/org/eclipse/emf/cdo/examples/server/offline/OfflineExampleMaster.java
That might at least explain why we had different results.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
| Re: [CDO] Multi master (clone) [message #932915 is a reply to message #931687] |
Thu, 04 October 2012 15:04 |
|
Am 03.10.2012 14:31, schrieb Ricky de Klerck:
> Eike,
>
> Thanks again for the quick reply. We managed to get our tests to work with the new files you gave me. We musst have
> had the wrong files after all..
>
> We're now testing the capabilities of the offline clone, but we're having some trouble right from the start..
> Our test:
>
> - start master on pc 1
> - start clone on pc 2
> - start clone on pc 3
> - start client connected to clone on pc 2
> - start client connected to clone on pc 3
>
> - create resource on pc 2
> - create object and save on pc 2
>
> (so far, so good: everything is synchronized)
>
> - disconnect pc 2 from network
> - create object and save on pc 2
> - (timeout exception occurs here)
> - connect pc 2 to network
>
> Result:
>
> - new object visible in master database
> - new object visible in clone database on pc 3
> - new object not visible in clone database on pc 2
>
> So it looks like the commit is done when the connecting is back, but the clone (pc 2) isn't synchronized. Any idea
> what the problem is? I suppose this test should work..
I think it should. Unfortunately I see not much in your description that could help me to get an idea about the cause
for your problem. Repository replication is one of the most complex setups in CDO. Doesn't the debugger help you to get
more details about the problem? Are there any exceptions?
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
| Re: [CDO] Multi master (clone) [message #936987 is a reply to message #936964] |
Mon, 08 October 2012 14:43 |
|
Am 08.10.2012 16:25, schrieb Ricky de Klerck:
>> I'm getting a timeout exception in my clone when I'm trying to save (network disconnected). After this exception the
>> state changes to offline. In my client I'm receiving a classnotfoundexception (TransportException). Could one of the
>> exceptions prevent the synchronizer from synchronize?
>
>> What exactly are these exceptions? Stack traces?
> Yes they are stack traces.
I know. Most of the time exceptions have stack traces. It's just incomprehensible to me why I'm supposed to guess how
they look like in order to help with your problem.
> But what about the classnotfound exception?
They can harm. The stack traces would help to judge that. Normally they indicate that a needed bundle is not properly
activated.
>
>> The normal way is that it keeps trying to commit till the connection is back online?
>
>> The client or the clone?
> Well, I assume the client? But, is it possible to keep working on the client (clone) when connection is lost?
Yes, that's the whole point of the offline clones. Of course not all functionality can work in oflline state. But the
fundamental functionality is available, including the abilities to read the entire object graph and commit changes.
> Now it looks like, the program is freezed, untill connection is restored?
If something appears frozen it's best to suspend the main thread in the debugger and find out what it's doing / waiting for.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
>> The clone doesn't commit to the master by itself. It continuously tries to connect > to the master in order to
>> receive replication updates.
>
> Ok, that's clear.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
| Re: [CDO] Multi master (clone) [message #937731 is a reply to message #937695] |
Tue, 09 October 2012 08:06 |
|
Am 09.10.2012 09:27, schrieb Ricky de Klerck:
> This is a part of the stack trace (I'm getting this in the client, if I save when network is disconnected):
>
>
> !ENTRY org.eclipse.net4j.util 4 0 2012-10-09 09:23:25.644
> !MESSAGE org.eclipse.emf.cdo.common.util.TransportException
> !STACK 0
> java.lang.ClassNotFoundException: org.eclipse.emf.cdo.common.util.TransportException
> at org.eclipse.osgi.internal.loader.BundleLoader.findClassInternal(BundleLoader.java:513)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:429)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:417)
> at org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.loadClass(DefaultClassLoader.java:107)
> at java.lang.ClassLoader.loadClass(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$ClassLoaderClassResolver.resolveClass(ExtendedIOUtil.java:284)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$3.resolveClass(ExtendedIOUtil.java:147)
> at java.io.ObjectInputStream.readNonProxyDesc(Unknown Source)
> at java.io.ObjectInputStream.readClassDesc(Unknown Source)
> at java.io.ObjectInputStream.readOrdinaryObject(Unknown Source)
> at java.io.ObjectInputStream.readObject0(Unknown Source)
> at java.io.ObjectInputStream.readObject(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:161)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:115)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.deserializeThrowable(RemoteExceptionIndication.java:76)
So the root cause is an exception on the server. The effect you're seeing is that this exception can not be deserialized
by the client, which, by itself, is not much of a problem because the root cause exception message and stack trace are
additionally transfered in text form, too.
I suggest that we first focus on the root cause exception on the server and then decide whether we want/can correctly
transfer it to the client by adding an Eclipse-Buddy registration to the bundle that contains the exception class (for
future debugging).
> at org.eclipse.net4j.signal.RemoteExceptionIndication.indicating(RemoteExceptionIndication.java:51)
> at org.eclipse.net4j.signal.Indication.doExtendedInput(Indication.java:57)
> at org.eclipse.net4j.signal.Signal.doInput(Signal.java:328)
> at org.eclipse.net4j.signal.Indication.execute(Indication.java:51)
> at org.eclipse.net4j.signal.Signal.runSync(Signal.java:253)
> at org.eclipse.net4j.signal.Signal.run(Signal.java:149)
> at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
> at java.lang.Thread.run(Unknown Source)
> [ERROR] org.eclipse.emf.cdo.common.util.CDOException
> java.lang.ClassNotFoundException: org.eclipse.emf.cdo.common.util.CDOException
> at org.eclipse.osgi.internal.loader.BundleLoader.findClassInternal(BundleLoader.java:513)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:429)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:417)
> at org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.loadClass(DefaultClassLoader.java:107)
> at java.lang.ClassLoader.loadClass(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$ClassLoaderClassResolver.resolveClass(ExtendedIOUtil.java:284)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$3.resolveClass(ExtendedIOUtil.java:147)
> at java.io.ObjectInputStream.readNonProxyDesc(Unknown Source)
> at java.io.ObjectInputStream.readClassDesc(Unknown Source)
> at java.io.ObjectInputStream.readNonProxyDesc(Unknown Source)
> at java.io.ObjectInputStream.readClassDesc(Unknown Source)
> at java.io.ObjectInputStream.readOrdinaryObject(Unknown Source)
> at java.io.ObjectInputStream.readObject0(Unknown Source)
> at java.io.ObjectInputStream.readObject(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:161)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:115)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.deserializeThrowable(RemoteExceptionIndication.java:76)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.indicating(RemoteExceptionIndication.java:51)
> at org.eclipse.net4j.signal.Indication.doExtendedInput(Indication.java:57)
> at org.eclipse.net4j.signal.Signal.doInput(Signal.java:328)
> at org.eclipse.net4j.signal.Indication.execute(Indication.java:51)
> at org.eclipse.net4j.signal.Signal.runSync(Signal.java:253)
> at org.eclipse.net4j.signal.Signal.run(Signal.java:149)
> at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
> at java.lang.Thread.run(Unknown Source)
>
> !ENTRY org.eclipse.net4j.util 4 0 2012-10-09 09:23:25.664
> !MESSAGE org.eclipse.emf.cdo.common.util.CDOException
> !STACK 0
> java.lang.ClassNotFoundException: org.eclipse.emf.cdo.common.util.CDOException
> at org.eclipse.osgi.internal.loader.BundleLoader.findClassInternal(BundleLoader.java:513)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:429)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:417)
> at org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.loadClass(DefaultClassLoader.java:107)
> at java.lang.ClassLoader.loadClass(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$ClassLoaderClassResolver.resolveClass(ExtendedIOUtil.java:284)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$3.resolveClass(ExtendedIOUtil.java:147)
> at java.io.ObjectInputStream.readNonProxyDesc(Unknown Source)
> at java.io.ObjectInputStream.readClassDesc(Unknown Source)
> at java.io.ObjectInputStream.readNonProxyDesc(Unknown Source)
> at java.io.ObjectInputStream.readClassDesc(Unknown Source)
> at java.io.ObjectInputStream.readOrdinaryObject(Unknown Source)
> at java.io.ObjectInputStream.readObject0(Unknown Source)
> at java.io.ObjectInputStream.readObject(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:161)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:115)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.deserializeThrowable(RemoteExceptionIndication.java:76)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.indicating(RemoteExceptionIndication.java:51)
> at org.eclipse.net4j.signal.Indication.doExtendedInput(Indication.java:57)
> at org.eclipse.net4j.signal.Signal.doInput(Signal.java:328)
> at org.eclipse.net4j.signal.Indication.execute(Indication.java:51)
> at org.eclipse.net4j.signal.Signal.runSync(Signal.java:253)
> at org.eclipse.net4j.signal.Signal.run(Signal.java:149)
> at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
> at java.lang.Thread.run(Unknown Source)
> [ERROR] org.eclipse.emf.cdo.common.util.TransportException
> java.lang.ClassNotFoundException: org.eclipse.emf.cdo.common.util.TransportException
> at org.eclipse.osgi.internal.loader.BundleLoader.findClassInternal(BundleLoader.java:513)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:429)
> at org.eclipse.osgi.internal.loader.BundleLoader.findClass(BundleLoader.java:417)
> at org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.loadClass(DefaultClassLoader.java:107)
> at java.lang.ClassLoader.loadClass(Unknown Source)
> at java.lang.Class.forName0(Native Method)
> at java.lang.Class.forName(Unknown Source)
> at java.io.ObjectInputStream.resolveClass(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil$3.resolveClass(ExtendedIOUtil.java:154)
> at java.io.ObjectInputStream.readNonProxyDesc(Unknown Source)
> at java.io.ObjectInputStream.readClassDesc(Unknown Source)
> at java.io.ObjectInputStream.readOrdinaryObject(Unknown Source)
> at java.io.ObjectInputStream.readObject0(Unknown Source)
> at java.io.ObjectInputStream.readObject(Unknown Source)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:161)
> at org.eclipse.net4j.util.io.ExtendedIOUtil.readObject(ExtendedIOUtil.java:115)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.deserializeThrowable(RemoteExceptionIndication.java:76)
> at org.eclipse.net4j.signal.RemoteExceptionIndication.indicating(RemoteExceptionIndication.java:51)
> at org.eclipse.net4j.signal.Indication.doExtendedInput(Indication.java:57)
> at org.eclipse.net4j.signal.Signal.doInput(Signal.java:328)
> at org.eclipse.net4j.signal.Indication.execute(Indication.java:51)
> at org.eclipse.net4j.signal.Signal.runSync(Signal.java:253)
> at org.eclipse.net4j.signal.Signal.run(Signal.java:149)
> at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
> at java.lang.Thread.run(Unknown Source)
>
>
> I tried adding org.eclipse.emf.cdo.common.util to the list of dependencies, but I can't find it! Has this something to
> do with the version of CDO I'm using (4.2)?
That's a package and it's contained by a bundle that's on the class path anyway. So PDE likely filters it from the
selection list.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
|
|
|
| Re: [CDO] Multi master (clone) [message #945051 is a reply to message #944998] |
Mon, 15 October 2012 08:04 |
|
Am 15.10.2012 09:37, schrieb Ricky de Klerck:
> Eike Stepper wrote on Tue, 09 October 2012 04:44
>> Am 09.10.2012 10:15, schrieb Ricky de Klerck:
>> > Well, the plugin org.eclipse.emf.cdo.common is selected in the run configuration. I suppose the package should be
>> in > there..
>> >
>> > When I do the same test with the example client it allso freezes and after a few seconds it asks if I want to retry
>> or > abort the operation.. I cannot do anything else at that moment.
>> I plan to spend some time on testing the offline replication mechanism. Hopefully by end of this week. But I can not
>> promise it 100%.
>
>
> Eike, any progress on this?
No, I'm sorry! And it seems more and more unlikely that I manage to get at this before end of the EclipseCon next week.
Still these possible issues with the offline mode are very high on my list.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
|
|
|
|
] 
Search
Help
Register
Login
Home
