« Return to Thread: Repast 2.0 with terracotta?

Re: Re past 2.0 with terracotta?

by Altaweel, Mark R. :: Rate this Message:

| View in Thread

Do you know at what point the run gets hung? Is there a way to debug and see where the code hangs?

There should be a 2D display that comes up in the demo. It should show agents moving around randomly more or less.

Mark
On Jan 5, 2011, at 7:05 PM, charlie1kimo wrote:

>
> Hi Mark,
> First thanks for the quick reply.
> I went through the stream of message in Feb 2009, and apparently the problem
> was not solved...
> I still have trouble running repast with terracotta in eclipse because of
> the strange path look up issues (like repast was run in
> repast.simphony.bin_and_src.jar when run with terracotta DSO, and was run in
> repast.simphony.runtime.bin when run without terracotta), but at least gave
> me some ideas and directions to try for.
>
> So I actually combine my run parameters from with and without terracotta,
> and run it from the command line. Now terracotta client starts okay, it
> connected to the sever, I can see the connection through terracotta develop
> console...etc. But when I make both master and worker to be localhost, and
> initialize the model, it's having the same problem as running without
> terracotta. It's the problem with stuck at the "starting workers" step.
> Here is the console output:
>
> 2011-01-05 16:58:29,409 INFO - Terracotta 3.4.0, as of 20101104-141154
> (Revision 16614 by cruise@su10mo4 from 3.4)
> 2011-01-05 16:58:29,804 INFO - Successfully loaded base configuration from
> file at
> '/data2/home/charlie/workspace/test_repast_terracotta_demo/testrepastterracottademo.rs/tc-config.xml'.
> 2011-01-05 16:58:29,995 INFO - Log file:
> '/data2/home/charlie/terracotta/client-logs/terracotta-client.log'.
> 2011-01-05 16:58:31,322 INFO - Connection successfully established to server
> at 127.0.0.1:9510
> Starting...:
> list with routing IDs         = [ localhost ]
> routing ID for fail-over node = localhost
> -- starting workers...
>
> There's something happened in client-log that lead to a CommonShutDownHook,
> which closed the local worker node I believed, but I am not sure what's the
> reason.
>
> I attached the server log and client log here for the reference
> http://old.nabble.com/file/p30601758/terracotta-client.log
> terracotta-client.log
> http://old.nabble.com/file/p30601758/terracotta-server.log
> terracotta-server.log
>
> Lastly, is this demo suppose to create display on its own? I tried to create
> a 2D display myself but it would crash the repast window...
>
> Thanks for the help!
>
> -Charlie-
>
>
> Altaweel, Mark R. wrote:
>>
>> There is a similar problem by another user back from February 2009.
>>
>> I listed the link to the stream of messages below:
>>
>> http://sourceforge.net/mailarchive/forum.php?thread_name=5FDF261785B44C4DB302979A658D006102AE3BA8%40spike.dis.anl.gov&forum_name=repast-interest
>>
>>
>> On Jan 4, 2011, at 8:40 PM, charlie1kimo wrote:
>>
>>>
>>> hmm...
>>> I believed it wouldn't be a firewall issue here.
>>> I could NOT run the demo with terracotta DSO because it was giving me the
>>> log4j error.
>>> The repast windows wouldn't start, and the console just showed log4j was
>>> not
>>> found and needed to be initialized.
>>>
>>> So I tried running the demo LOCALLY, which is just run the demo as a
>>> standard java/repast application.
>>> I put both my master and worker as "localhost," so it wouldn't have the
>>> firewall problem.
>>> Somehow, running locally on localhost made "starting workers" hang.
>>>
>>> I've also tried the terracotta's HelloClusteredWorld sample on our
>>> machine,
>>> and I can start the clients within eclipse and connect to the localhost
>>> server, which means terracotta is working...
>>>
>>> I guess for some reasons when eclipse runs terracotta DSO project, it
>>> missed
>>> the path to find log4j?
>>>
>>> Any ideas about what happened would be highly appreciated
>>>
>>> Thanks
>>>
>>> -Charlie-
>>>
>>>
>>> Altaweel, Mark R. wrote:
>>>>
>>>> One thing you might want to check is the firewall settings on your
>>>> remote
>>>> nodes. It sounds like some communication might be stalled some place and
>>>> this might be preventing terracota from communicating between your
>>>> nodes.
>>>> The demo should have agents basically moving around the grid; however,
>>>> outside of that there is not much behavior as this is just a demo of
>>>> using
>>>> terracotta.
>>>>
>>>> Also, if you might want to look at the repast.distributedBatch plugin
>>>> (more info see:
>>>> http://repast.sourceforge.net/docs/RepastParameterSweepsGettingStarted.pdf).
>>>> This enables distributed batch runs using multiple nodes, which might be
>>>> useful for the type of simulations you want to do.
>>>>
>>>> Mark
>>>> On Jan 4, 2011, at 6:50 PM, charlie1kimo wrote:
>>>>
>>>>>
>>>>> Hi all,
>>>>> Recently I am trying to run repast with terracotta as well. I tried to
>>>>> get
>>>>> the terracotta demo in repast package (the one with AgentManager.java
>>>>> and
>>>>> DoSomeActionAgent.java) to work but to no avail.
>>>>> Is this demo suppose to do something? or this demo is just a template
>>>>> for
>>>>> putting terracotta in your project?
>>>>>
>>>>> Here's my test runs with terracott:
>>>>> I've copied the terracotta demo into a separate project, and started
>>>>> from
>>>>> there.
>>>>> I configured the eclipse terracotta plugins and run configuration to
>>>>> read
>>>>> the correct tc-config.xml as well.
>>>>> Due to the bug in eclipse 64bit and terracotta plugin, I have to start
>>>>> the
>>>>> terracotta server from command line, but that's ok because the
>>>>> terracotta
>>>>> server started fine; I can also connect to the terracotta managing
>>>>> console.
>>>>>
>>>>> Then I ran this repast terracotta demo with DSO terracotta, and the
>>>>> following is what I got:
>>>>>
>>>>> 2011-01-04 16:41:33,160 INFO - Terracotta 3.4.0, as of 20101104-141154
>>>>> (Revision 16614 by cruise@su10mo4 from 3.4)
>>>>> 2011-01-04 16:41:33,566 INFO - Successfully loaded base configuration
>>>>> from
>>>>> file at
>>>>> '/data2/home/charlie/workspace/test_repast_terracotta_demo/testrepastterracottademo.rs/tc-config.xml'.
>>>>> 2011-01-04 16:41:33,753 INFO - Log file:
>>>>> '/data2/home/charlie/workspace/test_repast_terracotta_demo/terracotta/client-logs/terracotta-client.log'.
>>>>> 2011-01-04 16:41:35,126 INFO - Connection successfully established to
>>>>> server
>>>>> at 127.0.0.1:9510
>>>>> log4j:WARN No appenders could be found for logger
>>>>> (MessageCenter.INTERNAL.saf.core.runtime.Boot).
>>>>> log4j:WARN Please initialize the log4j system properly.
>>>>>
>>>>> Apparently it can connect to my server, but it died immediately after
>>>>> the
>>>>> connection due to the famous log4j error. (seems lots of people have
>>>>> this
>>>>> error)
>>>>>
>>>>> So I tried to run this demo without terracotta, and it seemed to work,
>>>>> except it hanged at "starting workers" when I initialized it:
>>>>>
>>>>> Starting...:
>>>>> list with routing IDs      = [ localhost ]
>>>>> routing ID for fail-over node = localhost
>>>>> -- starting workers...
>>>>>
>>>>> Is this the right behavior for this demo? I have a feeling that this
>>>>> demo
>>>>> is
>>>>> actually doing NOTHING.
>>>>> anyone has any ideas? In addition, I believed my log4j can be found in
>>>>> the
>>>>> build path, since it ran without terracotta successfully, it just
>>>>> hanged
>>>>> at
>>>>> the "starting workers."
>>>>>
>>>>> Lastly, some side notes for people who also testing this, when I ran
>>>>> this
>>>>> demo without terracotta DSO, I have to input parameter for "routerIDs."
>>>>> "routerIDs" has to have at least 2 hosts, separated by comma.
>>>>>
>>>>> Thanks in advance for any suggestions
>>>>>
>>>>> -Charlie-
>>>>>
>>>>> --
>>>>> View this message in context:
>>>>> http://old.nabble.com/Repast-2.0-with-terracotta--tp30592413p30592413.html
>>>>> Sent from the repast-interest mailing list archive at Nabble.com.
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Learn how Oracle Real Application Clusters (RAC) One Node allows
>>>>> customers
>>>>> to consolidate database storage, standardize their database
>>>>> environment,
>>>>> and,
>>>>> should the need arise, upgrade to a full multi-node Oracle RAC database
>>>>> without downtime or disruption
>>>>> http://p.sf.net/sfu/oracle-sfdevnl
>>>>> _______________________________________________
>>>>> Repast-interest mailing list
>>>>> Repast-interest@...
>>>>> https://lists.sourceforge.net/lists/listinfo/repast-interest
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Learn how Oracle Real Application Clusters (RAC) One Node allows
>>>> customers
>>>> to consolidate database storage, standardize their database environment,
>>>> and,
>>>> should the need arise, upgrade to a full multi-node Oracle RAC database
>>>> without downtime or disruption
>>>> http://p.sf.net/sfu/oracle-sfdevnl
>>>> _______________________________________________
>>>> Repast-interest mailing list
>>>> Repast-interest@...
>>>> https://lists.sourceforge.net/lists/listinfo/repast-interest
>>>>
>>>>
>>>
>>> --
>>> View this message in context:
>>> http://old.nabble.com/Repast-2.0-with-terracotta--tp30592413p30592919.html
>>> Sent from the repast-interest mailing list archive at Nabble.com.
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Learn how Oracle Real Application Clusters (RAC) One Node allows
>>> customers
>>> to consolidate database storage, standardize their database environment,
>>> and,
>>> should the need arise, upgrade to a full multi-node Oracle RAC database
>>> without downtime or disruption
>>> http://p.sf.net/sfu/oracle-sfdevnl
>>> _______________________________________________
>>> Repast-interest mailing list
>>> Repast-interest@...
>>> https://lists.sourceforge.net/lists/listinfo/repast-interest
>>
>>
>> ------------------------------------------------------------------------------
>> Learn how Oracle Real Application Clusters (RAC) One Node allows customers
>> to consolidate database storage, standardize their database environment,
>> and,
>> should the need arise, upgrade to a full multi-node Oracle RAC database
>> without downtime or disruption
>> http://p.sf.net/sfu/oracle-sfdevnl
>> _______________________________________________
>> Repast-interest mailing list
>> Repast-interest@...
>> https://lists.sourceforge.net/lists/listinfo/repast-interest
>>
>>
>
> --
> View this message in context: http://old.nabble.com/Repast-2.0-with-terracotta--tp30592413p30601758.html
> Sent from the repast-interest mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> Learn how Oracle Real Application Clusters (RAC) One Node allows customers
> to consolidate database storage, standardize their database environment, and,
> should the need arise, upgrade to a full multi-node Oracle RAC database
> without downtime or disruption
> http://p.sf.net/sfu/oracle-sfdevnl
> _______________________________________________
> Repast-interest mailing list
> Repast-interest@...
> https://lists.sourceforge.net/lists/listinfo/repast-interest


------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and,
should the need arise, upgrade to a full multi-node Oracle RAC database
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________
Repast-interest mailing list
Repast-interest@...
https://lists.sourceforge.net/lists/listinfo/repast-interest

 « Return to Thread: Repast 2.0 with terracotta?