You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jena.apache.org by Christoph Herzog <he...@ec.tuwien.ac.at> on 2011/10/25 15:12:30 UTC
Fuseki + TDB + Pellet -> corrupting data on INSERT
Hello,
I'm trying to get Fuseki to run with a TDB dataset and the pellet
reasoner
integrated.
That works pretty well so far, but I found that whenever I insert data
through
the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
data.
Everything works fine - reasoning and all - but when I stop the Fuseki
server
(by killing the process - is there another way?) it "breaks". Once I
start it
again and send any query to the SPARQL endpoint it replies with a HTTP
500 -
internal server error.
I found that I'm able to insert data with the tdbloader. In this case I
can
restart the server and it works normally. But also here, as soon as I
send an
INSERT statement it accepts it normally - but then when I shut down and
restart
the server I get the HTTP 500 error.
I'm pretty desperate allready, because inserting data into the server is
the
main task in my application...
This is the error message I get for any query if the server "breaks":
14:27:12 WARN Fuseki :: [1] RC = 500 : null
14:27:12 INFO Fuseki :: [1] 500 Server Error
The logger seems not to want to tell me more about it...
And this is (the relevant part of) my config file:
<#pelletGraph> rdf:type ja:RDFDataset ;
ja:defaultGraph <#reasonerModel> ;
tdb:unionDefaultGraph true ;
.
<#reasonerModel> rdf:type ja:InfModel ;
ja:reasoner [ja:reasonerClass
"org.mindswap.pellet.jena.PelletReasonerFactory" ;
] ;
ja:baseModel <#baseGraph> ;
.
<#baseGraph> rdf:type tdb:GraphTDB ;
tdb:location "dataset" ;
.
Please ask for anything else I you need to know (complete setup or the
like).
Any hints appreaciated!!
Thanks in advance,
Chris
Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Dave Reynolds <da...@gmail.com>.
On Tue, 2011-10-25 at 15:12 +0200, Christoph Herzog wrote:
Not directly related to your problem but I noticed:
> And this is (the relevant part of) my config file:
>
> <#pelletGraph> rdf:type ja:RDFDataset ;
> ja:defaultGraph <#reasonerModel> ;
> tdb:unionDefaultGraph true ;
I don't think you want that last line, though since the dataset is not a
TDBDataSet it will presumably make no difference.
BTW If you try without Pellet do you get the same corruption?
Dave
Re: [Fuseky] Form too large
Posted by Andy Seaborne <an...@apache.org>.
On 28/10/11 17:37, Damian Steer wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 28/10/11 17:14, Camel Christophe wrote:
>> Thanks for your help !
>>
>> Unfortunately, it didn't work and I found why: the maximum form
>> content size is set in the code of Fuseky ! In SPARQLServer.java
>> code source, there is the following line:
>>
>> // Set the max form size much higher.
>> context.setMaxFormContentSize(1*1024*1024) ;
>>
>> 1Mb is nice but not enough for me, knowing that I use Elda, a
>> linked-data implementation, that generates potentially very large
>> SPARQL queries.
You can reduce them by setting the endpoint to use SPARQL 1.1.
>
> (From Alex Dutton on IRC)
>
> "the max form size issue on the mailing list is the result of sending
> the query application/www-form-encoded, and it goes away if you send
> it as a query param (IIRC)
> I don't have access to an e-mail client that doesn't mangle
> In-Reply_To headers atm)"
Yes.
>
> I think Alex is suggesting POSTing multipart / form-data will work
> around this issue.
And in the development version of Fuseki:
1/ The default limit is 10M
2/ You can give it a jetty configuration file and do what you
want.(--jetty-config=...)
Andy
>
> Damian
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.11 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iEYEARECAAYFAk6q2lQACgkQAyLCB+mTtym/jgCguApgff3itWuDpl4yo6Nka0JE
> 67MAoNDXynQBAuXCU6kWehlvWlrjBRiV
> =2zXp
> -----END PGP SIGNATURE-----
Re: [Fuseky] Form too large
Posted by Damian Steer <d....@bristol.ac.uk>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 28/10/11 17:14, Camel Christophe wrote:
> Thanks for your help !
>
> Unfortunately, it didn't work and I found why: the maximum form
> content size is set in the code of Fuseky ! In SPARQLServer.java
> code source, there is the following line:
>
> // Set the max form size much higher.
> context.setMaxFormContentSize(1*1024*1024) ;
>
> 1Mb is nice but not enough for me, knowing that I use Elda, a
> linked-data implementation, that generates potentially very large
> SPARQL queries.
(From Alex Dutton on IRC)
"the max form size issue on the mailing list is the result of sending
the query application/www-form-encoded, and it goes away if you send
it as a query param (IIRC)
I don't have access to an e-mail client that doesn't mangle
In-Reply_To headers atm)"
I think Alex is suggesting POSTing multipart / form-data will work
around this issue.
Damian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAk6q2lQACgkQAyLCB+mTtym/jgCguApgff3itWuDpl4yo6Nka0JE
67MAoNDXynQBAuXCU6kWehlvWlrjBRiV
=2zXp
-----END PGP SIGNATURE-----
RE: [Fuseky] Form too large
Posted by Camel Christophe <cc...@cls.fr>.
Thanks for your help !
Unfortunately, it didn't work and I found why: the maximum form content size is set in the code of Fuseky !
In SPARQLServer.java code source, there is the following line:
// Set the max form size much higher.
context.setMaxFormContentSize(1*1024*1024) ;
1Mb is nice but not enough for me, knowing that I use Elda, a linked-data implementation, that generates potentially very large SPARQL queries.
Would there be a chance to make this configurable, through an assembler description or directly in the command line ?
Thanks,
Chris
-----Message d'origine-----
De : Damian Steer [mailto:d.steer@bristol.ac.uk]
Envoyé : vendredi 28 octobre 2011 18:01
À : jena-users@incubator.apache.org
Objet : Re: [Fuseky] Form too large
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 28/10/11 16:46, Camel Christophe wrote:
> Hello all !
>
> I'm using Fuseky 0.2.0 and I face an exception when submitting
> very large queries:
>
> java.lang.IllegalStateException: Form too large4618594>1048576
> I know the problem is related to Jetty, but I can't configure the
> Fuseky/Jetty couple for working with higher form size. Especially
> the " org.mortbay.jetty.Request.maxFormContentSize" system property
> does not work as expected.
I think the relevant setting is now (Jetty 7 onwards):
org.eclipse.jetty.server.Request.maxFormContentSize
See if that works,
Damian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAk6q0aYACgkQAyLCB+mTtymY1gCdFTf/QPfUZHtFFwSLSoj+moDK
LMMAoJ4a4VqIlybAVoX0ACVnf9h1RNAb
=x0Gx
-----END PGP SIGNATURE-----
Re: [Fuseky] Form too large
Posted by Damian Steer <d....@bristol.ac.uk>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 28/10/11 16:46, Camel Christophe wrote:
> Hello all !
>
> I'm using Fuseky 0.2.0 and I face an exception when submitting
> very large queries:
>
> java.lang.IllegalStateException: Form too large4618594>1048576
> I know the problem is related to Jetty, but I can't configure the
> Fuseky/Jetty couple for working with higher form size. Especially
> the " org.mortbay.jetty.Request.maxFormContentSize" system property
> does not work as expected.
I think the relevant setting is now (Jetty 7 onwards):
org.eclipse.jetty.server.Request.maxFormContentSize
See if that works,
Damian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAk6q0aYACgkQAyLCB+mTtymY1gCdFTf/QPfUZHtFFwSLSoj+moDK
LMMAoJ4a4VqIlybAVoX0ACVnf9h1RNAb
=x0Gx
-----END PGP SIGNATURE-----
[Fuseky] Form too large
Posted by Camel Christophe <cc...@cls.fr>.
Hello all !
I'm using Fuseky 0.2.0 and I face an exception when submitting very
large queries:
java.lang.IllegalStateException: Form too large4618594>1048576
at
org.eclipse.jetty.server.Request.extractParameters(Request.java:251)
at
org.eclipse.jetty.server.Request.getParameter(Request.java:659)
at
org.openjena.fuseki.servlets.SPARQL_Query.validate(SPARQL_Query.java:167
)
I know the problem is related to Jetty, but I can't configure the
Fuseky/Jetty couple for working with higher form size. Especially the "
org.mortbay.jetty.Request.maxFormContentSize" system property does not
work as expected.
Is there a simple way to get around the limitation Fuseky in Jetty ?
Many thanks,
Chris
Correction: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Christoph Herzog <he...@ec.tuwien.ac.at>.
Hello Andy, hello all,
I think I messed up the upload, the ttl was from an (fruitless) experiment.
I re-uploaded the correct version to:
http://www.megafileupload.com/en/file/331059/fuseki-tdb-pellet-data-corrupted-on-insert-reupload-zip.html
But anyway, it is exactly the same with the following ttl:
(fuseki-start with "--desc pellet.ttl /dataset" accordingly)
=== === === === === ===
@prefix fuseki: <http://jena.apache.org/fuseki#> .
@prefix tdb: <http://jena.hpl.hp.com/2008/tdb#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix ja: <http://jena.hpl.hp.com/2005/11/Assembler#> .
@prefix : <#> .
# Declaration additional assembler items.
[] ja:loadClass "com.hp.hpl.jena.tdb.TDB" .
# TDB
tdb:DatasetTDB rdfs:subClassOf ja:RDFDataset .
tdb:GraphTDB rdfs:subClassOf ja:Model .
# DATA SECTION
<#pelletGraph> rdf:type ja:RDFDataset ;
ja:defaultGraph <#reasonerModel> ;
.
<#reasonerModel> rdf:type ja:InfModel ;
ja:reasoner [ja:reasonerClass
"org.mindswap.pellet.jena.PelletReasonerFactory" ;
] ;
ja:baseModel <#baseGraph> ;
.
<#baseGraph> rdf:type tdb:GraphTDB ;
tdb:location "dataset" ;
.
=== === === === === ===
Note that the fuseki-init script also should load the pizza.owl file from the ontology directory (wrong filename).
Sorry for messing this up :-(
I also tried it completely without pellet - in this case I do not get a Server Error.
Although with this ttl I noticed that the updates (inserts) I did are no longer present when I restart the server:
=== snipplet ===
<#pelletGraph> rdf:type ja:RDFDataset ;
ja:defaultGraph <#baseGraph> ;
.
<#baseGraph> rdf:type tdb:GraphTDB ;
tdb:location "dataset" ;
.
=== === ===
The Java version, btw, is:
> java version "1.6.0_24"
> Java(TM) SE Runtime Environment (build 1.6.0_24-b07)
> Java HotSpot(TM) 64-Bit Server VM (build 19.1-b02, mixed mode)
Please ask for any further details. I can also re-upload the corrected versions
Thanks to all,
Chris
-----Ursprüngliche Nachricht-----
Von: Andy Seaborne [mailto:andy.seaborne.apache@gmail.com] Im Auftrag von Andy Seaborne
Gesendet: Sonntag, 30. Oktober 2011 17:59
An: jena-users@incubator.apache.org
Betreff: Re: Details: Fuseki + TDB + Pellet -> corrupting data on INSERT
On 28/10/11 15:33, Christoph Herzog wrote:
You have Pellet running over "dataset" but you directly insert into
"dataset" without going via Pellet.
Could you try doing this without Pellet at all - at the moment, I can't
see what's going on whatsoever (e.g. I don't run Windows nor Pellet).
(What does "java -version" say? 32 bit or 64 bit?)
Andy
Re: Details: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Andy Seaborne <an...@apache.org>.
On 28/10/11 15:33, Christoph Herzog wrote:
> Hello again,
> I tried a few variations with the setup and also with other ontologies
> and didn't get any improvement.
>
> I think in order to best describe the problem it's best to show the
> complete setup. I uploaded it to:
>
> http://www.megafileupload.com/en/file/330870/fuseki-tdb-pellet-insert-pr
> oblem-zip.html
> [File: fuseki_tdb_pellet_insert_problem.zip, 32 MB]
>
> Inside you find the fuseki server and the scripts I party generated
> myself. Working on a windows system I use the .bat files. I created
> equivalent shell scripts but haven't tested them.
>
> What I do is:
> A) Initiate the TDB with the fuseki-init.bat script
> -> This loads the pizza.owl ontology from the ontology directory into
> TDB (directory "dataset")
> B) Start Fuseki using the fuseki-start.bat script
> -> All works well
>
> Now I go to the fuseki web frontend (localhost:3030), Control Panel,
> dataset:
> 1) I enter the test query in the ontology directory (select-test.sparql)
> it returns 4 pizzas - which is correct.
> 2) I can now restart fuseki and repeat step 1 - it still works
> 3) Now I enter the insert query from the ontology directory on the web
> frontend (insert-test.sparql) - accepted.
> 4) I enter the test query from step 1 again and now it finds 5 pizzas as
> expected.
> 5) I restart fuseki and try the test query again - I get a 500 server
> error.
>
>
> Maybe someone finds the problem in my setup?
You have Pellet running over "dataset" but you directly insert into
"dataset" without going via Pellet.
Could you try doing this without Pellet at all - at the moment, I can't
see what's going on whatsoever (e.g. I don't run Windows nor Pellet).
(What does "java -version" say? 32 bit or 64 bit?)
Andy
>
> Thanks in advance to all the helpful people on this list!
> Chris
Details: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Christoph Herzog <he...@ec.tuwien.ac.at>.
Hello again,
I tried a few variations with the setup and also with other ontologies
and didn't get any improvement.
I think in order to best describe the problem it's best to show the
complete setup. I uploaded it to:
http://www.megafileupload.com/en/file/330870/fuseki-tdb-pellet-insert-pr
oblem-zip.html
[File: fuseki_tdb_pellet_insert_problem.zip, 32 MB]
Inside you find the fuseki server and the scripts I party generated
myself. Working on a windows system I use the .bat files. I created
equivalent shell scripts but haven't tested them.
What I do is:
A) Initiate the TDB with the fuseki-init.bat script
-> This loads the pizza.owl ontology from the ontology directory into
TDB (directory "dataset")
B) Start Fuseki using the fuseki-start.bat script
-> All works well
Now I go to the fuseki web frontend (localhost:3030), Control Panel,
dataset:
1) I enter the test query in the ontology directory (select-test.sparql)
it returns 4 pizzas - which is correct.
2) I can now restart fuseki and repeat step 1 - it still works
3) Now I enter the insert query from the ontology directory on the web
frontend (insert-test.sparql) - accepted.
4) I enter the test query from step 1 again and now it finds 5 pizzas as
expected.
5) I restart fuseki and try the test query again - I get a 500 server
error.
Maybe someone finds the problem in my setup?
Thanks in advance to all the helpful people on this list!
Chris
Re: AW: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Andy Seaborne <an...@apache.org>.
On 25/10/11 16:31, Christoph Herzog wrote:
> No, it appears the unwanted behaviour persists even when I fall back to Fuseki 0.2.0 (with TDB 0.8.10).
> I fear it is probably not an artefact of the new transaction features :-(
>
> Best,
> Chris
You are not using the new transaction features - Fuseki does not use
them yet.
It's impossible to tell what's going on -
>> That works pretty well so far, but I found that whenever I insert data
>> through
>> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
>> data.
Are you posting data through Pellet? If you add data underneath Pellet
then any caches Pellet has are invalid but it isn't gong to notice.
>> I found that I'm able to insert data with the tdbloader. In this case I
>> can
>> restart the server and it works normally. But also here, as soon as I
>> send an
>> INSERT statement it accepts it normally - but then when I shut down and
>> restart
>> the server I get the HTTP 500 error.
>>
>> I'm pretty desperate allready, because inserting data into the server is
>> the
>> main task in my application...
>>
>> This is the error message I get for any query if the server "breaks":
>> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
>> 14:27:12 INFO Fuseki :: [1] 500 Server Error
More details? Is this during startup? Is it the first request? What
is the query? Or update?
Andy
>
>
> -----Ursprüngliche Nachricht-----
> Von: Paolo Castagna [mailto:castagna.lists@googlemail.com]
> Gesendet: Dienstag, 25. Oktober 2011 16:45
> An: jena-users@incubator.apache.org
> Betreff: Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
>
> Hi Christoph,
> which version of Fuseki (and TDB) are you using?
>
> There have been some changes in TDB recently (to add support for
> transactions) and there might be problems if you use -SNAPSHOTs
> or trunk.
>
> Paolo
>
> Christoph Herzog wrote:
>> Hello,
>> I'm trying to get Fuseki to run with a TDB dataset and the pellet
>> reasoner
>> integrated.
>>
>> That works pretty well so far, but I found that whenever I insert data
>> through
>> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
>> data.
>>
>> Everything works fine - reasoning and all - but when I stop the Fuseki
>> server
>> (by killing the process - is there another way?) it "breaks". Once I
>> start it
>> again and send any query to the SPARQL endpoint it replies with a HTTP
>> 500 -
>> internal server error.
>>
>> I found that I'm able to insert data with the tdbloader. In this case I
>> can
>> restart the server and it works normally. But also here, as soon as I
>> send an
>> INSERT statement it accepts it normally - but then when I shut down and
>> restart
>> the server I get the HTTP 500 error.
>>
>> I'm pretty desperate allready, because inserting data into the server is
>> the
>> main task in my application...
>>
>> This is the error message I get for any query if the server "breaks":
>> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
>> 14:27:12 INFO Fuseki :: [1] 500 Server Error
>>
>> The logger seems not to want to tell me more about it...
>>
>> And this is (the relevant part of) my config file:
>>
>> <#pelletGraph> rdf:type ja:RDFDataset ;
>> ja:defaultGraph<#reasonerModel> ;
>> tdb:unionDefaultGraph true ;
>> .
>>
>> <#reasonerModel> rdf:type ja:InfModel ;
>> ja:reasoner [ja:reasonerClass
>> "org.mindswap.pellet.jena.PelletReasonerFactory" ;
>> ] ;
>> ja:baseModel<#baseGraph> ;
>> .
>>
>> <#baseGraph> rdf:type tdb:GraphTDB ;
>> tdb:location "dataset" ;
>> .
>>
>>
>>
>> Please ask for anything else I you need to know (complete setup or the
>> like).
>> Any hints appreaciated!!
>>
>>
>>
>> Thanks in advance,
>>
>> Chris
>>
>>
>
Re: AW: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Paolo Castagna <ca...@googlemail.com>.
Christoph Herzog wrote:
> No, it appears the unwanted behaviour persists even when I fall back to Fuseki 0.2.0 (with TDB 0.8.10).
> I fear it is probably not an artefact of the new transaction features :-(
TDB 0.8.10 should be fine, no changes related to transaction support.
The more details you give us on exactly what you are doing the sooner
we can find if there is a problem (and fix it).
If you describe exactly what you are doing, I can try to replicate and
see if I have your same issue. Command lines even without explanation
are great for me. :-)
Paolo
>
> Best,
> Chris
>
>
> -----Ursprüngliche Nachricht-----
> Von: Paolo Castagna [mailto:castagna.lists@googlemail.com]
> Gesendet: Dienstag, 25. Oktober 2011 16:45
> An: jena-users@incubator.apache.org
> Betreff: Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
>
> Hi Christoph,
> which version of Fuseki (and TDB) are you using?
>
> There have been some changes in TDB recently (to add support for
> transactions) and there might be problems if you use -SNAPSHOTs
> or trunk.
>
> Paolo
>
> Christoph Herzog wrote:
>> Hello,
>> I'm trying to get Fuseki to run with a TDB dataset and the pellet
>> reasoner
>> integrated.
>>
>> That works pretty well so far, but I found that whenever I insert data
>> through
>> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
>> data.
>>
>> Everything works fine - reasoning and all - but when I stop the Fuseki
>> server
>> (by killing the process - is there another way?) it "breaks". Once I
>> start it
>> again and send any query to the SPARQL endpoint it replies with a HTTP
>> 500 -
>> internal server error.
>>
>> I found that I'm able to insert data with the tdbloader. In this case I
>> can
>> restart the server and it works normally. But also here, as soon as I
>> send an
>> INSERT statement it accepts it normally - but then when I shut down and
>> restart
>> the server I get the HTTP 500 error.
>>
>> I'm pretty desperate allready, because inserting data into the server is
>> the
>> main task in my application...
>>
>> This is the error message I get for any query if the server "breaks":
>> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
>> 14:27:12 INFO Fuseki :: [1] 500 Server Error
>>
>> The logger seems not to want to tell me more about it...
>>
>> And this is (the relevant part of) my config file:
>>
>> <#pelletGraph> rdf:type ja:RDFDataset ;
>> ja:defaultGraph <#reasonerModel> ;
>> tdb:unionDefaultGraph true ;
>> .
>>
>> <#reasonerModel> rdf:type ja:InfModel ;
>> ja:reasoner [ja:reasonerClass
>> "org.mindswap.pellet.jena.PelletReasonerFactory" ;
>> ] ;
>> ja:baseModel <#baseGraph> ;
>> .
>>
>> <#baseGraph> rdf:type tdb:GraphTDB ;
>> tdb:location "dataset" ;
>> .
>>
>>
>>
>> Please ask for anything else I you need to know (complete setup or the
>> like).
>> Any hints appreaciated!!
>>
>>
>>
>> Thanks in advance,
>>
>> Chris
>>
>>
>
AW: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Christoph Herzog <he...@ec.tuwien.ac.at>.
No, it appears the unwanted behaviour persists even when I fall back to Fuseki 0.2.0 (with TDB 0.8.10).
I fear it is probably not an artefact of the new transaction features :-(
Best,
Chris
-----Ursprüngliche Nachricht-----
Von: Paolo Castagna [mailto:castagna.lists@googlemail.com]
Gesendet: Dienstag, 25. Oktober 2011 16:45
An: jena-users@incubator.apache.org
Betreff: Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
Hi Christoph,
which version of Fuseki (and TDB) are you using?
There have been some changes in TDB recently (to add support for
transactions) and there might be problems if you use -SNAPSHOTs
or trunk.
Paolo
Christoph Herzog wrote:
> Hello,
> I'm trying to get Fuseki to run with a TDB dataset and the pellet
> reasoner
> integrated.
>
> That works pretty well so far, but I found that whenever I insert data
> through
> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
> data.
>
> Everything works fine - reasoning and all - but when I stop the Fuseki
> server
> (by killing the process - is there another way?) it "breaks". Once I
> start it
> again and send any query to the SPARQL endpoint it replies with a HTTP
> 500 -
> internal server error.
>
> I found that I'm able to insert data with the tdbloader. In this case I
> can
> restart the server and it works normally. But also here, as soon as I
> send an
> INSERT statement it accepts it normally - but then when I shut down and
> restart
> the server I get the HTTP 500 error.
>
> I'm pretty desperate allready, because inserting data into the server is
> the
> main task in my application...
>
> This is the error message I get for any query if the server "breaks":
> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
> 14:27:12 INFO Fuseki :: [1] 500 Server Error
>
> The logger seems not to want to tell me more about it...
>
> And this is (the relevant part of) my config file:
>
> <#pelletGraph> rdf:type ja:RDFDataset ;
> ja:defaultGraph <#reasonerModel> ;
> tdb:unionDefaultGraph true ;
> .
>
> <#reasonerModel> rdf:type ja:InfModel ;
> ja:reasoner [ja:reasonerClass
> "org.mindswap.pellet.jena.PelletReasonerFactory" ;
> ] ;
> ja:baseModel <#baseGraph> ;
> .
>
> <#baseGraph> rdf:type tdb:GraphTDB ;
> tdb:location "dataset" ;
> .
>
>
>
> Please ask for anything else I you need to know (complete setup or the
> like).
> Any hints appreaciated!!
>
>
>
> Thanks in advance,
>
> Chris
>
>
Re: AW: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Paolo Castagna <ca...@googlemail.com>.
Christoph Herzog wrote:
> Hello Paolo,
> I'm using Fuseki version 0.2.1-SNAPSHOT (actually 0.2.1-20110915.071835-17).
> The TDB is delivered with Fuseki and from the POM.xml I think it is TDB 0.8.11-SNAPSHOT.
> I'll try replacing the TDB (or falling back to Fuseki 0.2.0) - which version of TDB do you recommend?
Hi Chris,
the latest stable release of TDB is 0.8.10:
http://repo1.maven.org/maven2/com/hp/hpl/jena/tdb/0.8.10/
... if you find a bug here you *might* get a faster response.
If you are ok with being on the bleeding-edge and you want to
help us testing the latest TDB SNAPSHOTs, you can find them here:
https://repository.apache.org/content/repositories/snapshots/org/apache/jena/jena-tdb/0.9.0-incubating-SNAPSHOT/
I warn you: this is unreleased software and it might have problems.
You can find the list of issues related to TDB here:
http://bit.ly/tJtpKs
Paolo
>
> Thanks and best regards,
> Chris
>
>
> -----Ursprüngliche Nachricht-----
> Von: Paolo Castagna [mailto:castagna.lists@googlemail.com]
> Gesendet: Dienstag, 25. Oktober 2011 16:45
> An: jena-users@incubator.apache.org
> Betreff: Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
>
> Hi Christoph,
> which version of Fuseki (and TDB) are you using?
>
> There have been some changes in TDB recently (to add support for
> transactions) and there might be problems if you use -SNAPSHOTs
> or trunk.
>
> Paolo
>
> Christoph Herzog wrote:
>> Hello,
>> I'm trying to get Fuseki to run with a TDB dataset and the pellet
>> reasoner
>> integrated.
>>
>> That works pretty well so far, but I found that whenever I insert data
>> through
>> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
>> data.
>>
>> Everything works fine - reasoning and all - but when I stop the Fuseki
>> server
>> (by killing the process - is there another way?) it "breaks". Once I
>> start it
>> again and send any query to the SPARQL endpoint it replies with a HTTP
>> 500 -
>> internal server error.
>>
>> I found that I'm able to insert data with the tdbloader. In this case I
>> can
>> restart the server and it works normally. But also here, as soon as I
>> send an
>> INSERT statement it accepts it normally - but then when I shut down and
>> restart
>> the server I get the HTTP 500 error.
>>
>> I'm pretty desperate allready, because inserting data into the server is
>> the
>> main task in my application...
>>
>> This is the error message I get for any query if the server "breaks":
>> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
>> 14:27:12 INFO Fuseki :: [1] 500 Server Error
>>
>> The logger seems not to want to tell me more about it...
>>
>> And this is (the relevant part of) my config file:
>>
>> <#pelletGraph> rdf:type ja:RDFDataset ;
>> ja:defaultGraph <#reasonerModel> ;
>> tdb:unionDefaultGraph true ;
>> .
>>
>> <#reasonerModel> rdf:type ja:InfModel ;
>> ja:reasoner [ja:reasonerClass
>> "org.mindswap.pellet.jena.PelletReasonerFactory" ;
>> ] ;
>> ja:baseModel <#baseGraph> ;
>> .
>>
>> <#baseGraph> rdf:type tdb:GraphTDB ;
>> tdb:location "dataset" ;
>> .
>>
>>
>>
>> Please ask for anything else I you need to know (complete setup or the
>> like).
>> Any hints appreaciated!!
>>
>>
>>
>> Thanks in advance,
>>
>> Chris
>>
>>
>
AW: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Christoph Herzog <he...@ec.tuwien.ac.at>.
Hello Paolo,
I'm using Fuseki version 0.2.1-SNAPSHOT (actually 0.2.1-20110915.071835-17).
The TDB is delivered with Fuseki and from the POM.xml I think it is TDB 0.8.11-SNAPSHOT.
I'll try replacing the TDB (or falling back to Fuseki 0.2.0) - which version of TDB do you recommend?
Thanks and best regards,
Chris
-----Ursprüngliche Nachricht-----
Von: Paolo Castagna [mailto:castagna.lists@googlemail.com]
Gesendet: Dienstag, 25. Oktober 2011 16:45
An: jena-users@incubator.apache.org
Betreff: Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
Hi Christoph,
which version of Fuseki (and TDB) are you using?
There have been some changes in TDB recently (to add support for
transactions) and there might be problems if you use -SNAPSHOTs
or trunk.
Paolo
Christoph Herzog wrote:
> Hello,
> I'm trying to get Fuseki to run with a TDB dataset and the pellet
> reasoner
> integrated.
>
> That works pretty well so far, but I found that whenever I insert data
> through
> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
> data.
>
> Everything works fine - reasoning and all - but when I stop the Fuseki
> server
> (by killing the process - is there another way?) it "breaks". Once I
> start it
> again and send any query to the SPARQL endpoint it replies with a HTTP
> 500 -
> internal server error.
>
> I found that I'm able to insert data with the tdbloader. In this case I
> can
> restart the server and it works normally. But also here, as soon as I
> send an
> INSERT statement it accepts it normally - but then when I shut down and
> restart
> the server I get the HTTP 500 error.
>
> I'm pretty desperate allready, because inserting data into the server is
> the
> main task in my application...
>
> This is the error message I get for any query if the server "breaks":
> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
> 14:27:12 INFO Fuseki :: [1] 500 Server Error
>
> The logger seems not to want to tell me more about it...
>
> And this is (the relevant part of) my config file:
>
> <#pelletGraph> rdf:type ja:RDFDataset ;
> ja:defaultGraph <#reasonerModel> ;
> tdb:unionDefaultGraph true ;
> .
>
> <#reasonerModel> rdf:type ja:InfModel ;
> ja:reasoner [ja:reasonerClass
> "org.mindswap.pellet.jena.PelletReasonerFactory" ;
> ] ;
> ja:baseModel <#baseGraph> ;
> .
>
> <#baseGraph> rdf:type tdb:GraphTDB ;
> tdb:location "dataset" ;
> .
>
>
>
> Please ask for anything else I you need to know (complete setup or the
> like).
> Any hints appreaciated!!
>
>
>
> Thanks in advance,
>
> Chris
>
>
Re: Fuseki + TDB + Pellet -> corrupting data on INSERT
Posted by Paolo Castagna <ca...@googlemail.com>.
Hi Christoph,
which version of Fuseki (and TDB) are you using?
There have been some changes in TDB recently (to add support for
transactions) and there might be problems if you use -SNAPSHOTs
or trunk.
Paolo
Christoph Herzog wrote:
> Hello,
> I'm trying to get Fuseki to run with a TDB dataset and the pellet
> reasoner
> integrated.
>
> That works pretty well so far, but I found that whenever I insert data
> through
> the SPARQL endpoint or POST my ontology it would somehow "corrupt" the
> data.
>
> Everything works fine - reasoning and all - but when I stop the Fuseki
> server
> (by killing the process - is there another way?) it "breaks". Once I
> start it
> again and send any query to the SPARQL endpoint it replies with a HTTP
> 500 -
> internal server error.
>
> I found that I'm able to insert data with the tdbloader. In this case I
> can
> restart the server and it works normally. But also here, as soon as I
> send an
> INSERT statement it accepts it normally - but then when I shut down and
> restart
> the server I get the HTTP 500 error.
>
> I'm pretty desperate allready, because inserting data into the server is
> the
> main task in my application...
>
> This is the error message I get for any query if the server "breaks":
> 14:27:12 WARN Fuseki :: [1] RC = 500 : null
> 14:27:12 INFO Fuseki :: [1] 500 Server Error
>
> The logger seems not to want to tell me more about it...
>
> And this is (the relevant part of) my config file:
>
> <#pelletGraph> rdf:type ja:RDFDataset ;
> ja:defaultGraph <#reasonerModel> ;
> tdb:unionDefaultGraph true ;
> .
>
> <#reasonerModel> rdf:type ja:InfModel ;
> ja:reasoner [ja:reasonerClass
> "org.mindswap.pellet.jena.PelletReasonerFactory" ;
> ] ;
> ja:baseModel <#baseGraph> ;
> .
>
> <#baseGraph> rdf:type tdb:GraphTDB ;
> tdb:location "dataset" ;
> .
>
>
>
> Please ask for anything else I you need to know (complete setup or the
> like).
> Any hints appreaciated!!
>
>
>
> Thanks in advance,
>
> Chris
>
>