You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@stanbol.apache.org by Andrea Taurchini <at...@gmail.com> on 2012/11/15 14:36:11 UTC
Stopping the framework ...
Dear All,
maybe I'm missing (again) something, but if I stop the framework, no matter
if through Felix Web Console or CTRL+C, configurations go to hell on the
next restart.
Even the default enhancement chain will stop working since the order or the
engine is changed to :
- *metaxa* ( optional , currently not available)
- *entityhubExtraction* ( required , currently not available)
- *tika* ( optional , TikaEngine)
- *langdetect* ( required , LanguageDetectionEnhancementEngine)
- *ner* ( required , NamedEntityExtractionEnhancementEngine)
- *dbpediaLinking* ( required , NamedEntityTaggingEngine)
not to mention the fact that my own configurations (topic classifier ...)
is completely removed ...
Thanks,
Andrea
Re: Stopping the framework ...
Posted by Rupert Westenthaler <ru...@gmail.com>.
Hi,
even after a detailed inspection of the log file you provided was not
able to find an indication of any problem. Based on the logging the
Stanbol instance was started, stopped and than started and stopped
again. The loggings of the first and the second startup are really
similar.
The only thing that might cause problems is the "java.io.IOException:
Unable to establish loopback connection" but as you mentioned in your
last mail solving this has also not solved your issue.
On Mon, Nov 19, 2012 at 4:17 PM, Andrea Taurchini <at...@gmail.com> wrote:
> I should install stanbol on a windows server ... so it is not possible ?
AFAIK there are several users that do use Stanbol on Windows. Only
yesterday a Blog about MakoLab using Stanbol with their Windows CMS
was posted [1].
Andrea can you try to do the following
1. 1st time start of Stanbol (in an empty directory)
2. after the start archive the "stanbol\config" folder
3. stop the stanbol instance
4. after shutdown again archive the "stanbol\config" folder
5. start stanbol a 2nd time
6. make an third archive of the "stanbol\config" folder
If you can send me those three archives I will make and before after
check of the OSGI component configurations as written to your HD.
Maybe this will provide a hint about your issue
best
Rupert
[1] http://blog.iks-project.eu/makolabs-stanbol-integration-with-renault-international-cms-system/
--
| Rupert Westenthaler rupert.westenthaler@gmail.com
| Bodenlehenstraße 11 ++43-699-11108907
| A-5500 Bischofshofen
Re: Stopping the framework ...
Posted by Andrea Taurchini <at...@gmail.com>.
Dear Rupert,
unfortunately yes I do have the same error.
I tried to solve as said in the link you provided ... but again the same
error.
I should install stanbol on a windows server ... so it is not possible ?
Thanks.
Best,
Andrea
2012/11/19 Rupert Westenthaler <ru...@gmail.com>
> Hi,
>
> do you still have the same issue?
>
> I noticed in the logs
>
> 19.11.2012 11:00:53.701 *ERROR* [Jetty HTTP Service]
> org.apache.felix.http.jetty Exception while initializing Jetty.
> (java.io.IOException: Unable to establish loopback connection)
> java.io.IOException: Unable to establish loopback connection
>
> This indicates that you are using Windows with a Firewall/Virus
> scanner configuration that blocks the java non-blockin IO. I have
> already written about that on the Stanbol lists
>
> http://markmail.org/message/6iiagoqpbklbgxw6
>
> hope this helps
> best
> Rupert
>
> On Mon, Nov 19, 2012 at 2:31 PM, Andrea Taurchini <at...@gmail.com>
> wrote:
> > Dear Rupert,
> > thanks for your help.
> >
> > Best,
> > Andrea
> >
> >
> >
> > 2012/11/19 Rupert Westenthaler <ru...@gmail.com>
> >>
> >> Hi,
> >>
> >> you can not send attachments via the list. Feel free to send it directly
> >> to me.
> >>
> >> On Mon, Nov 19, 2012 at 11:09 AM, Andrea Taurchini <
> ataurchini@gmail.com>
> >> wrote:
> >> > 1) clean stanbol folder
> >> > 2) launch java -Xmx1g -jar -XX:MaxPermSize=128m
> >> >
> >> >
> stanbol_src\launchers\full\target\org.apache.stanbol.launchers.full-0.10.0-SNAPSHOT.jar
> >>
> >> -XX:MaxPermSize=128m is not enough for the full launcher as it
> >> requires ~200 MByte. You should use -XX:MaxPermSize=256m instead.
> >>
> >> With only 128m of PermGen memory I would expect the full launcher to
> >> throw OutOfMemory exceptions during startup. If this happens the
> >> initial configuration (created during the first startup) will be
> >> incomplete and corrupted. This could indeed explain the issues you are
> >> experiences.
> >>
> >> best
> >> Rupert
> >>
> >>
> >> > 3) once fully active ... stop the service with CTRL^C
> >> > 4) wait for stopping the service
> >> > 5) relaunch the same startup command
> >> > 6) verify that entityhubExtraction enhancer is no more available
> >> >
> >> > Thanks for your help.
> >> >
> >> >
> >> > Best,
> >> > Andrea
> >> >
> >> >
> >> >
> >> >
> >> >
> >> >
> >> > 2012/11/15 Rupert Westenthaler <ru...@gmail.com>
> >> >>
> >> >> Hi
> >> >>
> >> >> On Thu, Nov 15, 2012 at 2:36 PM, Andrea Taurchini
> >> >> <at...@gmail.com>
> >> >> wrote:
> >> >> > Dear All,
> >> >> > maybe I'm missing (again) something, but if I stop the framework,
> no
> >> >> > matter
> >> >> > if through Felix Web Console or CTRL+C, configurations go to hell
> on
> >> >> > the
> >> >> > next restart.
> >> >>
> >> >> No you are missing nothing. All those ways to shutdown Stanbol should
> >> >> work just fine. I can not remember having ever a problem like that.
> >> >>
> >> >> > Even the default enhancement chain will stop working since the
> order
> >> >> > or
> >> >> > the
> >> >> > engine is changed to :
> >> >> >
> >> >> > - *metaxa* ( optional , currently not available)
> >> >> > - *entityhubExtraction* ( required , currently not available)
> >> >> > - *tika* ( optional , TikaEngine)
> >> >> > - *langdetect* ( required , LanguageDetectionEnhancementEngine)
> >> >> > - *ner* ( required , NamedEntityExtractionEnhancementEngine)
> >> >> > - *dbpediaLinking* ( required , NamedEntityTaggingEngine)
> >> >> >
> >> >>
> >> >> that "not available" engines are listed first is expected for the
> >> >> WeightedChain. This chain determines the order based on information
> >> >> provided by the Engine. So if an Engine is not available such
> >> >> Information are not available. As the order does not matter for
> >> >> Engines that are not available my decision was to list them first.
> >> >>
> >> >> > not to mention the fact that my own configurations (topic
> classifier
> >> >> > ...)
> >> >> > is completely removed ...
> >> >> >
> >> >>
> >> >> Somehow it looks like as OSGI is not able to write files to the disc.
> >> >> Can you please check the Stanbol log file
> >> >> {launcher-dir}/stanbol/logs/error.log if you can find related
> >> >> information.
> >> >>
> >> >> best
> >> >> Rupert
> >> >>
> >> >> > Thanks,
> >> >> > Andrea
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> | Rupert Westenthaler rupert.westenthaler@gmail.com
> >> >> | Bodenlehenstraße 11 ++43-699-11108907
> >> >> | A-5500 Bischofshofen
> >> >
> >> >
> >>
> >>
> >>
> >> --
> >> | Rupert Westenthaler rupert.westenthaler@gmail.com
> >> | Bodenlehenstraße 11 ++43-699-11108907
> >> | A-5500 Bischofshofen
> >
> >
>
>
>
> --
> | Rupert Westenthaler rupert.westenthaler@gmail.com
> | Bodenlehenstraße 11 ++43-699-11108907
> | A-5500 Bischofshofen
>
Re: Stopping the framework ...
Posted by Rupert Westenthaler <ru...@gmail.com>.
Hi,
you can not send attachments via the list. Feel free to send it directly to me.
On Mon, Nov 19, 2012 at 11:09 AM, Andrea Taurchini <at...@gmail.com> wrote:
> 1) clean stanbol folder
> 2) launch java -Xmx1g -jar -XX:MaxPermSize=128m
> stanbol_src\launchers\full\target\org.apache.stanbol.launchers.full-0.10.0-SNAPSHOT.jar
-XX:MaxPermSize=128m is not enough for the full launcher as it
requires ~200 MByte. You should use -XX:MaxPermSize=256m instead.
With only 128m of PermGen memory I would expect the full launcher to
throw OutOfMemory exceptions during startup. If this happens the
initial configuration (created during the first startup) will be
incomplete and corrupted. This could indeed explain the issues you are
experiences.
best
Rupert
> 3) once fully active ... stop the service with CTRL^C
> 4) wait for stopping the service
> 5) relaunch the same startup command
> 6) verify that entityhubExtraction enhancer is no more available
>
> Thanks for your help.
>
>
> Best,
> Andrea
>
>
>
>
>
>
> 2012/11/15 Rupert Westenthaler <ru...@gmail.com>
>>
>> Hi
>>
>> On Thu, Nov 15, 2012 at 2:36 PM, Andrea Taurchini <at...@gmail.com>
>> wrote:
>> > Dear All,
>> > maybe I'm missing (again) something, but if I stop the framework, no
>> > matter
>> > if through Felix Web Console or CTRL+C, configurations go to hell on the
>> > next restart.
>>
>> No you are missing nothing. All those ways to shutdown Stanbol should
>> work just fine. I can not remember having ever a problem like that.
>>
>> > Even the default enhancement chain will stop working since the order or
>> > the
>> > engine is changed to :
>> >
>> > - *metaxa* ( optional , currently not available)
>> > - *entityhubExtraction* ( required , currently not available)
>> > - *tika* ( optional , TikaEngine)
>> > - *langdetect* ( required , LanguageDetectionEnhancementEngine)
>> > - *ner* ( required , NamedEntityExtractionEnhancementEngine)
>> > - *dbpediaLinking* ( required , NamedEntityTaggingEngine)
>> >
>>
>> that "not available" engines are listed first is expected for the
>> WeightedChain. This chain determines the order based on information
>> provided by the Engine. So if an Engine is not available such
>> Information are not available. As the order does not matter for
>> Engines that are not available my decision was to list them first.
>>
>> > not to mention the fact that my own configurations (topic classifier
>> > ...)
>> > is completely removed ...
>> >
>>
>> Somehow it looks like as OSGI is not able to write files to the disc.
>> Can you please check the Stanbol log file
>> {launcher-dir}/stanbol/logs/error.log if you can find related
>> information.
>>
>> best
>> Rupert
>>
>> > Thanks,
>> > Andrea
>>
>>
>>
>> --
>> | Rupert Westenthaler rupert.westenthaler@gmail.com
>> | Bodenlehenstraße 11 ++43-699-11108907
>> | A-5500 Bischofshofen
>
>
--
| Rupert Westenthaler rupert.westenthaler@gmail.com
| Bodenlehenstraße 11 ++43-699-11108907
| A-5500 Bischofshofen
Re: Stopping the framework ...
Posted by Andrea Taurchini <at...@gmail.com>.
Dear Rupert and all,
unfortunately I can't see any error related to disk writing.
Here attached you can find the error log can you please help finding which
is the problem ?
I made the following :
1) clean stanbol folder
2) launch java -Xmx1g -jar -XX:MaxPermSize=128m
stanbol_src\launchers\full\target\org.apache.stanbol.launchers.full-0.10.0-SNAPSHOT.jar
3) once fully active ... stop the service with CTRL^C
4) wait for stopping the service
5) relaunch the same startup command
6) verify that entityhubExtraction enhancer is no more available
Thanks for your help.
Best,
Andrea
2012/11/15 Rupert Westenthaler <ru...@gmail.com>
> Hi
>
> On Thu, Nov 15, 2012 at 2:36 PM, Andrea Taurchini <at...@gmail.com>
> wrote:
> > Dear All,
> > maybe I'm missing (again) something, but if I stop the framework, no
> matter
> > if through Felix Web Console or CTRL+C, configurations go to hell on the
> > next restart.
>
> No you are missing nothing. All those ways to shutdown Stanbol should
> work just fine. I can not remember having ever a problem like that.
>
> > Even the default enhancement chain will stop working since the order or
> the
> > engine is changed to :
> >
> > - *metaxa* ( optional , currently not available)
> > - *entityhubExtraction* ( required , currently not available)
> > - *tika* ( optional , TikaEngine)
> > - *langdetect* ( required , LanguageDetectionEnhancementEngine)
> > - *ner* ( required , NamedEntityExtractionEnhancementEngine)
> > - *dbpediaLinking* ( required , NamedEntityTaggingEngine)
> >
>
> that "not available" engines are listed first is expected for the
> WeightedChain. This chain determines the order based on information
> provided by the Engine. So if an Engine is not available such
> Information are not available. As the order does not matter for
> Engines that are not available my decision was to list them first.
>
> > not to mention the fact that my own configurations (topic classifier ...)
> > is completely removed ...
> >
>
> Somehow it looks like as OSGI is not able to write files to the disc.
> Can you please check the Stanbol log file
> {launcher-dir}/stanbol/logs/error.log if you can find related
> information.
>
> best
> Rupert
>
> > Thanks,
> > Andrea
>
>
>
> --
> | Rupert Westenthaler rupert.westenthaler@gmail.com
> | Bodenlehenstraße 11 ++43-699-11108907
> | A-5500 Bischofshofen
>
Re: Stopping the framework ...
Posted by Rupert Westenthaler <ru...@gmail.com>.
Hi
On Thu, Nov 15, 2012 at 2:36 PM, Andrea Taurchini <at...@gmail.com> wrote:
> Dear All,
> maybe I'm missing (again) something, but if I stop the framework, no matter
> if through Felix Web Console or CTRL+C, configurations go to hell on the
> next restart.
No you are missing nothing. All those ways to shutdown Stanbol should
work just fine. I can not remember having ever a problem like that.
> Even the default enhancement chain will stop working since the order or the
> engine is changed to :
>
> - *metaxa* ( optional , currently not available)
> - *entityhubExtraction* ( required , currently not available)
> - *tika* ( optional , TikaEngine)
> - *langdetect* ( required , LanguageDetectionEnhancementEngine)
> - *ner* ( required , NamedEntityExtractionEnhancementEngine)
> - *dbpediaLinking* ( required , NamedEntityTaggingEngine)
>
that "not available" engines are listed first is expected for the
WeightedChain. This chain determines the order based on information
provided by the Engine. So if an Engine is not available such
Information are not available. As the order does not matter for
Engines that are not available my decision was to list them first.
> not to mention the fact that my own configurations (topic classifier ...)
> is completely removed ...
>
Somehow it looks like as OSGI is not able to write files to the disc.
Can you please check the Stanbol log file
{launcher-dir}/stanbol/logs/error.log if you can find related
information.
best
Rupert
> Thanks,
> Andrea
--
| Rupert Westenthaler rupert.westenthaler@gmail.com
| Bodenlehenstraße 11 ++43-699-11108907
| A-5500 Bischofshofen