Home » Modeling » EMF » [CDO] Timeout while getting worker Thread?!
| [CDO] Timeout while getting worker Thread?! [message #486627] |
Fri, 18 September 2009 12:46  |
 Stefan Winkler
Messages: 307
Registered: July 2009
Location: Germany |
Senior Member |
|
|
Hi,
as some of you may already know, I am doing quite large calculations
which end up producing a lot of objects.
E.g., I have a calculation, which creates a resource with about
1,500,000 new CDO objects each of which have bidirectional references
to several existing CDO objects. Thus I get very large transactions.
I'm currently running into a weird timeout error which I don't really
know what to do about:
ERROR] java.util.concurrent.TimeoutException: Worker thread activation
timed out after 10000 millis
org.eclipse.net4j.util.lifecycle.LifecycleException:
java.util.concurrent.TimeoutException: Worker thread activation timed
out after 10000 millis
at
org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
at
org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
at
org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
at
org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
$CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
at
org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
at
org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
at
org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
.... several more ...
Caused by: java.util.concurrent.TimeoutException: Worker thread
activation timed out after 10000 millis
at org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
at
org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
at
org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
... 57 more
What could cause this timeout? Are there too few Worker Threads in the
pool? But then again, the pool should grow, shouldn't it?
Any ideas?
Cheers,
Stefan
|
|
|
| Re: [CDO] Timeout while getting worker Thread?! [message #486653 is a reply to message #486627] |
Fri, 18 September 2009 14:18  |
|
Hi Stefan,
Looks as if your box is just too slow :P
Can you reproduce this scenario in a test case?
We might factor the creation of the QueueRunner into a protected method
so that you can override it and set a higher activation timeout. Please
file a bugzilla.
Cheers
/Eike
----
http://thegordian.blogspot.com
Stefan Winkler schrieb:
> Hi,
>
> as some of you may already know, I am doing quite large calculations
> which end up producing a lot of objects.
>
> E.g., I have a calculation, which creates a resource with about
> 1,500,000 new CDO objects each of which have bidirectional references
> to several existing CDO objects. Thus I get very large transactions.
>
> I'm currently running into a weird timeout error which I don't really
> know what to do about:
>
> ERROR] java.util.concurrent.TimeoutException: Worker thread activation
> timed out after 10000 millis
> org.eclipse.net4j.util.lifecycle.LifecycleException:
> java.util.concurrent.TimeoutException: Worker thread activation timed
> out after 10000 millis
> at
> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
> at
> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
> at
> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
> at
> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
> at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
> $CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
> at
> org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
> at
> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
> at
> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
> .... several more ...
> Caused by: java.util.concurrent.TimeoutException: Worker thread
> activation timed out after 10000 millis
> at org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
> at
> org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
> at
> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
> ... 57 more
>
> What could cause this timeout? Are there too few Worker Threads in the
> pool? But then again, the pool should grow, shouldn't it?
>
> Any ideas?
>
> Cheers,
> Stefan
>
>
>
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
| Re: [CDO] Timeout while getting worker Thread?! [message #486677 is a reply to message #486653] |
Fri, 18 September 2009 14:35 |
Stefan Winkler
Messages: 307
Registered: July 2009
Location: Germany |
Senior Member |
|
|
Am Freitag, den 18.09.2009, 10:18 -0400 schrieb Eike Stepper:
> Hi Stefan,
>
> Looks as if your box is just too slow :P
Newly purchased Phenom II X4 945 @3GHz w/ 4GB RAM?
No way :-P
> Can you reproduce this scenario in a test case?
>
> We might factor the creation of the QueueRunner into a protected method
> so that you can override it and set a higher activation timeout. Please
> file a bugzilla.
I'm just doing a new run doing some intermediate commits every 100,000
objects. Maybe this helps. So, currently, I hope, I can't rproduce
it :-P
Cheers,
Stefan
> Cheers
> /Eike
>
> ----
> http://thegordian.blogspot.com
>
>
>
> Stefan Winkler schrieb:
> > Hi,
> >
> > as some of you may already know, I am doing quite large calculations
> > which end up producing a lot of objects.
> >
> > E.g., I have a calculation, which creates a resource with about
> > 1,500,000 new CDO objects each of which have bidirectional references
> > to several existing CDO objects. Thus I get very large transactions.
> >
> > I'm currently running into a weird timeout error which I don't really
> > know what to do about:
> >
> > ERROR] java.util.concurrent.TimeoutException: Worker thread activation
> > timed out after 10000 millis
> > org.eclipse.net4j.util.lifecycle.LifecycleException:
> > java.util.concurrent.TimeoutException: Worker thread activation timed
> > out after 10000 millis
> > at
> > org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
> > at
> > org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
> > at
> > org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
> > at
> > org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
> > at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
> > $CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
> > at
> > org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
> > at
> > org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
> > at
> > org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
> > .... several more ...
> > Caused by: java.util.concurrent.TimeoutException: Worker thread
> > activation timed out after 10000 millis
> > at org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
> > at
> > org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
> > at
> > org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
> > ... 57 more
> >
> > What could cause this timeout? Are there too few Worker Threads in the
> > pool? But then again, the pool should grow, shouldn't it?
> >
> > Any ideas?
> >
> > Cheers,
> > Stefan
> >
> >
> >
|
|
|
| Re: [CDO] Timeout while getting worker Thread?! [message #486818 is a reply to message #486677] |
Sat, 19 September 2009 13:30 |
Andre Dietisheim
Messages: 131
Registered: July 2009 |
Senior Member |
|
|
Hi Stefan
I experienced exactly the same when I wrote my concurrent tests for the
CDORevisionCaches. I did not track the problem down to its source but it
was clearly related to the number of concurrent accesses to the
session/transaction. So I unfortunately only can confirm without being
able to provide a solution ;-(
Cheers
André
Stefan Winkler wrote:
> Am Freitag, den 18.09.2009, 10:18 -0400 schrieb Eike Stepper:
>> Hi Stefan,
>>
>> Looks as if your box is just too slow :P
> Newly purchased Phenom II X4 945 @3GHz w/ 4GB RAM?
> No way :-P
>
>> Can you reproduce this scenario in a test case?
>>
>> We might factor the creation of the QueueRunner into a protected method
>> so that you can override it and set a higher activation timeout. Please
>> file a bugzilla.
>
> I'm just doing a new run doing some intermediate commits every 100,000
> objects. Maybe this helps. So, currently, I hope, I can't rproduce
> it :-P
>
> Cheers,
> Stefan
>
>
>> Cheers
>> /Eike
>>
>> ----
>> http://thegordian.blogspot.com
>>
>>
>>
>> Stefan Winkler schrieb:
>>> Hi,
>>>
>>> as some of you may already know, I am doing quite large calculations
>>> which end up producing a lot of objects.
>>>
>>> E.g., I have a calculation, which creates a resource with about
>>> 1,500,000 new CDO objects each of which have bidirectional references
>>> to several existing CDO objects. Thus I get very large transactions.
>>>
>>> I'm currently running into a weird timeout error which I don't really
>>> know what to do about:
>>>
>>> ERROR] java.util.concurrent.TimeoutException: Worker thread activation
>>> timed out after 10000 millis
>>> org.eclipse.net4j.util.lifecycle.LifecycleException:
>>> java.util.concurrent.TimeoutException: Worker thread activation timed
>>> out after 10000 millis
>>> at
>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
>>> at
>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
>>> at
>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
>>> at
>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
>>> at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
>>> $CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
>>> at
>>> org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
>>> at
>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
>>> at
>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
>>> .... several more ...
>>> Caused by: java.util.concurrent.TimeoutException: Worker thread
>>> activation timed out after 10000 millis
>>> at org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
>>> at
>>> org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
>>> at
>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
>>> ... 57 more
>>>
>>> What could cause this timeout? Are there too few Worker Threads in the
>>> pool? But then again, the pool should grow, shouldn't it?
>>>
>>> Any ideas?
>>>
>>> Cheers,
>>> Stefan
>>>
>>>
>>>
>
|
|
|
| Re: [CDO] Timeout while getting worker Thread?! [message #486819 is a reply to message #486818] |
Sat, 19 September 2009 13:42 |
|
Guys,
I still think that it's only related to processing power vs. number of
threads and their CPU usage. Being able to control the activation
timeout should fix it. Maybe we should generally increase the default
activation timeout?
Cheers
/Eike
----
http://thegordian.blogspot.com
Ande Dietisheim schrieb:
> Hi Stefan
>
> I experienced exactly the same when I wrote my concurrent tests for the
> CDORevisionCaches. I did not track the problem down to its source but it
> was clearly related to the number of concurrent accesses to the
> session/transaction. So I unfortunately only can confirm without being
> able to provide a solution ;-(
>
> Cheers
> André
>
> Stefan Winkler wrote:
>
>> Am Freitag, den 18.09.2009, 10:18 -0400 schrieb Eike Stepper:
>>
>>> Hi Stefan,
>>>
>>> Looks as if your box is just too slow :P
>>>
>> Newly purchased Phenom II X4 945 @3GHz w/ 4GB RAM?
>> No way :-P
>>
>>
>>> Can you reproduce this scenario in a test case?
>>>
>>> We might factor the creation of the QueueRunner into a protected method
>>> so that you can override it and set a higher activation timeout. Please
>>> file a bugzilla.
>>>
>> I'm just doing a new run doing some intermediate commits every 100,000
>> objects. Maybe this helps. So, currently, I hope, I can't rproduce
>> it :-P
>>
>> Cheers,
>> Stefan
>>
>>
>>
>>> Cheers
>>> /Eike
>>>
>>> ----
>>> http://thegordian.blogspot.com
>>>
>>>
>>>
>>> Stefan Winkler schrieb:
>>>
>>>> Hi,
>>>>
>>>> as some of you may already know, I am doing quite large calculations
>>>> which end up producing a lot of objects.
>>>>
>>>> E.g., I have a calculation, which creates a resource with about
>>>> 1,500,000 new CDO objects each of which have bidirectional references
>>>> to several existing CDO objects. Thus I get very large transactions.
>>>>
>>>> I'm currently running into a weird timeout error which I don't really
>>>> know what to do about:
>>>>
>>>> ERROR] java.util.concurrent.TimeoutException: Worker thread activation
>>>> timed out after 10000 millis
>>>> org.eclipse.net4j.util.lifecycle.LifecycleException:
>>>> java.util.concurrent.TimeoutException: Worker thread activation timed
>>>> out after 10000 millis
>>>> at
>>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
>>>> at
>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
>>>> at
>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
>>>> at
>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
>>>> at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
>>>> $CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
>>>> at
>>>> org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
>>>> at
>>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
>>>> at
>>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
>>>> .... several more ...
>>>> Caused by: java.util.concurrent.TimeoutException: Worker thread
>>>> activation timed out after 10000 millis
>>>> at org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
>>>> at
>>>> org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
>>>> at
>>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
>>>> ... 57 more
>>>>
>>>> What could cause this timeout? Are there too few Worker Threads in the
>>>> pool? But then again, the pool should grow, shouldn't it?
>>>>
>>>> Any ideas?
>>>>
>>>> Cheers,
>>>> Stefan
>>>>
>>>>
>>>>
>>>>
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
| Re: [CDO] Timeout while getting worker Thread?! [message #486954 is a reply to message #486819] |
Mon, 21 September 2009 09:47 |
Stefan Winkler
Messages: 307
Registered: July 2009
Location: Germany |
Senior Member |
|
|
Hi,
Comments below.
> Guys,
>
> I still think that it's only related to processing power vs. number of
> threads and their CPU usage. Being able to control the activation
> timeout should fix it. Maybe we should generally increase the default
> activation timeout?
I don't know for sure, but currently, I get a lot of OutOfMemoryErrors
(heap), even though my heap is 2,5G. As I said, I'm currently working on
getting my transactions smaller, maybe that helps. But what I wanted to
say is that it might even be that in the process of getting a worker
there could have been an OOME (even though it has not been reported)...
Regarding timeouts: I think what would help more than generally
increasing default timeouts would be documentation about which timeouts
exist and a facility to configure them. Because I start hitting
different 10s-timeouts now and again, even though I have set the
CDOSession-options-protocol timeout to "-1".
Cheers,
Stefan
>
> Cheers
> /Eike
>
> ----
> http://thegordian.blogspot.com
>
>
>
> Ande Dietisheim schrieb:
>> Hi Stefan
>>
>> I experienced exactly the same when I wrote my concurrent tests for the
>> CDORevisionCaches. I did not track the problem down to its source but it
>> was clearly related to the number of concurrent accesses to the
>> session/transaction. So I unfortunately only can confirm without being
>> able to provide a solution ;-(
>>
>> Cheers
>> André
>>
>> Stefan Winkler wrote:
>>
>>> Am Freitag, den 18.09.2009, 10:18 -0400 schrieb Eike Stepper:
>>>
>>>> Hi Stefan,
>>>>
>>>> Looks as if your box is just too slow :P
>>>>
>>> Newly purchased Phenom II X4 945 @3GHz w/ 4GB RAM?
>>> No way :-P
>>>
>>>
>>>> Can you reproduce this scenario in a test case?
>>>>
>>>> We might factor the creation of the QueueRunner into a protected
>>>> method so that you can override it and set a higher activation
>>>> timeout. Please file a bugzilla.
>>>>
>>> I'm just doing a new run doing some intermediate commits every 100,000
>>> objects. Maybe this helps. So, currently, I hope, I can't rproduce
>>> it :-P
>>>
>>> Cheers,
>>> Stefan
>>>
>>>
>>>
>>>> Cheers
>>>> /Eike
>>>>
>>>> ----
>>>> http://thegordian.blogspot.com
>>>>
>>>>
>>>>
>>>> Stefan Winkler schrieb:
>>>>
>>>>> Hi,
>>>>>
>>>>> as some of you may already know, I am doing quite large calculations
>>>>> which end up producing a lot of objects.
>>>>>
>>>>> E.g., I have a calculation, which creates a resource with about
>>>>> 1,500,000 new CDO objects each of which have bidirectional references
>>>>> to several existing CDO objects. Thus I get very large transactions.
>>>>>
>>>>> I'm currently running into a weird timeout error which I don't really
>>>>> know what to do about:
>>>>>
>>>>> ERROR] java.util.concurrent.TimeoutException: Worker thread
>>>>> activation
>>>>> timed out after 10000 millis
>>>>> org.eclipse.net4j.util.lifecycle.LifecycleException:
>>>>> java.util.concurrent.TimeoutException: Worker thread activation timed
>>>>> out after 10000 millis
>>>>> at
>>>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
>>>>>
>>>>> at
>>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
>>>>>
>>>>> at
>>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
>>>>>
>>>>> at
>>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
>>>>>
>>>>> at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
>>>>> $CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
>>>>> at
>>>>> org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
>>>>>
>>>>> at
>>>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
>>>>>
>>>>> at
>>>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
>>>>>
>>>>> .... several more ...
>>>>> Caused by: java.util.concurrent.TimeoutException: Worker thread
>>>>> activation timed out after 10000 millis
>>>>> at
>>>>> org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
>>>>> at
>>>>> org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
>>>>>
>>>>> at
>>>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
>>>>>
>>>>> ... 57 more
>>>>>
>>>>> What could cause this timeout? Are there too few Worker Threads in
>>>>> the
>>>>> pool? But then again, the pool should grow, shouldn't it?
>>>>> Any ideas?
>>>>>
>>>>> Cheers,
>>>>> Stefan
>>>>>
>>>>>
>>>>>
|
|
|
| Re: [CDO] Timeout while getting worker Thread?! [message #486956 is a reply to message #486954] |
Mon, 21 September 2009 10:00 |
|
Hi Stefan,
I agree, we should make as many of the timeouts as easily configurable
as possible. Instead of (or in addition to) offering to override the
code, we should add more options API. Can you file a bugzilla?
Cheers
/Eike
----
http://thegordian.blogspot.com
http://twitter.com/eikestepper
Stefan Winkler schrieb:
> Hi,
>
> Comments below.
>
>> Guys,
>>
>> I still think that it's only related to processing power vs. number of
>> threads and their CPU usage. Being able to control the activation
>> timeout should fix it. Maybe we should generally increase the default
>> activation timeout?
>>
> I don't know for sure, but currently, I get a lot of OutOfMemoryErrors
> (heap), even though my heap is 2,5G. As I said, I'm currently working on
> getting my transactions smaller, maybe that helps. But what I wanted to
> say is that it might even be that in the process of getting a worker
> there could have been an OOME (even though it has not been reported)...
>
> Regarding timeouts: I think what would help more than generally
> increasing default timeouts would be documentation about which timeouts
> exist and a facility to configure them. Because I start hitting
> different 10s-timeouts now and again, even though I have set the
> CDOSession-options-protocol timeout to "-1".
>
> Cheers,
> Stefan
>
>
>
>> Cheers
>> /Eike
>>
>> ----
>> http://thegordian.blogspot.com
>>
>>
>>
>> Ande Dietisheim schrieb:
>>
>>> Hi Stefan
>>>
>>> I experienced exactly the same when I wrote my concurrent tests for the
>>> CDORevisionCaches. I did not track the problem down to its source but it
>>> was clearly related to the number of concurrent accesses to the
>>> session/transaction. So I unfortunately only can confirm without being
>>> able to provide a solution ;-(
>>>
>>> Cheers
>>> André
>>>
>>> Stefan Winkler wrote:
>>>
>>>
>>>> Am Freitag, den 18.09.2009, 10:18 -0400 schrieb Eike Stepper:
>>>>
>>>>
>>>>> Hi Stefan,
>>>>>
>>>>> Looks as if your box is just too slow :P
>>>>>
>>>>>
>>>> Newly purchased Phenom II X4 945 @3GHz w/ 4GB RAM?
>>>> No way :-P
>>>>
>>>>
>>>>
>>>>> Can you reproduce this scenario in a test case?
>>>>>
>>>>> We might factor the creation of the QueueRunner into a protected
>>>>> method so that you can override it and set a higher activation
>>>>> timeout. Please file a bugzilla.
>>>>>
>>>>>
>>>> I'm just doing a new run doing some intermediate commits every 100,000
>>>> objects. Maybe this helps. So, currently, I hope, I can't rproduce
>>>> it :-P
>>>>
>>>> Cheers,
>>>> Stefan
>>>>
>>>>
>>>>
>>>>
>>>>> Cheers
>>>>> /Eike
>>>>>
>>>>> ----
>>>>> http://thegordian.blogspot.com
>>>>>
>>>>>
>>>>>
>>>>> Stefan Winkler schrieb:
>>>>>
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> as some of you may already know, I am doing quite large calculations
>>>>>> which end up producing a lot of objects.
>>>>>>
>>>>>> E.g., I have a calculation, which creates a resource with about
>>>>>> 1,500,000 new CDO objects each of which have bidirectional references
>>>>>> to several existing CDO objects. Thus I get very large transactions.
>>>>>>
>>>>>> I'm currently running into a weird timeout error which I don't really
>>>>>> know what to do about:
>>>>>>
>>>>>> ERROR] java.util.concurrent.TimeoutException: Worker thread
>>>>>> activation
>>>>>> timed out after 10000 millis
>>>>>> org.eclipse.net4j.util.lifecycle.LifecycleException:
>>>>>> java.util.concurrent.TimeoutException: Worker thread activation timed
>>>>>> out after 10000 millis
>>>>>> at
>>>>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:93)
>>>>>>
>>>>>> at
>>>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.getInval idationRunner(CDOSessionImpl.java:603)
>>>>>>
>>>>>> at
>>>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:541)
>>>>>>
>>>>>> at
>>>>>> org.eclipse.emf.internal.cdo.session.CDOSessionImpl.handleCo mmitNotification(CDOSessionImpl.java:473)
>>>>>>
>>>>>> at org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl
>>>>>> $CDOCommitContextImpl.postCommit(CDOTransactionImpl.java:134 7)
>>>>>> at
>>>>>> org.eclipse.emf.internal.cdo.transaction.CDOSingleTransactio nStrategyImpl.commit(CDOSingleTransactionStrategyImpl.java:7 2)
>>>>>>
>>>>>> at
>>>>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:598)
>>>>>>
>>>>>> at
>>>>>> org.eclipse.emf.internal.cdo.transaction.CDOTransactionImpl. commit(CDOTransactionImpl.java:612)
>>>>>>
>>>>>> .... several more ...
>>>>>> Caused by: java.util.concurrent.TimeoutException: Worker thread
>>>>>> activation timed out after 10000 millis
>>>>>> at
>>>>>> org.eclipse.net4j.util.concurrent.Worker.doActivate(Worker.j ava:94)
>>>>>> at
>>>>>> org.eclipse.net4j.util.concurrent.QueueWorker.doActivate(Que ueWorker.java:80)
>>>>>>
>>>>>> at
>>>>>> org.eclipse.net4j.util.lifecycle.Lifecycle.activate(Lifecycl e.java:65)
>>>>>>
>>>>>> ... 57 more
>>>>>>
>>>>>> What could cause this timeout? Are there too few Worker Threads in
>>>>>> the
>>>>>> pool? But then again, the pool should grow, shouldn't it?
>>>>>> Any ideas?
>>>>>>
>>>>>> Cheers,
>>>>>> Stefan
>>>>>>
>>>>>>
>>>>>>
>>>>>>
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
| |
Goto Forum:
Current Time: Thu Apr 25 07:35:32 GMT 2024
Powered by FUDForum. Page generated in 0.02737 seconds |
] 
Search
Help
Register
Login
Home
