Hi Joseph, Your logs do show that in the first case "talismane-nlp" is currently processing. In the 2nd example "EDFAcronymeLinking" and "EDFImplantationLinking" are in progress. So those loggings would suggest that the execution of the chain is indeed not finished after 60sec.
In case it would be a bug of the Chain Execution (EventJobManager) I would rather expect exceptions before the timeout or a timeout but with all engines are marked as completed. As it is very unlikely that processing of short text takes so much time this smells a lot like a bug, but as it is likely not in the chain execution itself it should be related to the EnhancementEngines used in the chain. The best is if you use the Stanbol Enhancer Stress Test Tool [1] with a "Plain Text" test file containing real test examples of your use case. Usually it is sufficient to create ~50 unique texts and then duplicate those in the file until you get more examples as you would like to configure to "stanbol.it.multithreadtest.requests". The tools will allow you the steadily increase the number of threads. It also prints statistics about the time required by the different EnhancementEngines to process the texts (min, max, average processing time). So by doing so it should be possible to acquire more information on what is going. Would be nice if you could perform such tests and report back with results. best Rupert [1] https://stanbol.apache.org/docs/trunk/utils/enhancerstresstest On Sun, Jun 16, 2013 at 7:21 PM, Joseph M'Bimbi-Bene <[email protected]> wrote: > Hello everybody, i'm again having problems with Stanbol. I try to extract > named entities from a lot of documents. I get a lot of timeouts from > linking engines or nlp processing, it happens very early in the process, > after like the 10th document and then happens almost systematically. here > are some logs. > > > Can someone tell me what could cause these timeouts and what can i do to > prevent them. > > > The texts i try to enhance are very short, and the process is > multithreaded, there may be up to 10 threads making calls to Stanbol, can > it be the cause ? > > > (org.apache.stanbol.enhancer.servicesapi.ChainException: Execution timeout > after 60.001sec (timeout:60sec) for ContentItem > <urn:content-item-sha1-f53cb5608144eb1503239ed1c1143ab92f867677> > > Status: > > -EDFImplantationLinking: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusSheduled >> > > -langdetect: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusCompleted >> > > -EDFAcronymeLinking: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusSheduled >> > > -talismane-nlp: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusInProgress >> > > To change the timeout change value of property > 'stanbol.maxEnhancementJobWaitTime' for the service class > org.apache.stanbol.enhancer.jobmanager.event.impl.EventJobManagerImpl) > org.apache.stanbol.enhancer.servicesapi.ChainException: Execution timeout > after 60.001sec (timeout:60sec) for ContentItem > > ------------------ > > 06.06.2013 19:12:10.512 *WARN* [14651473@qtp-8593864-312] > org.apache.felix.http.jetty /enhancer/chain/EDFEnhancement > (org.apache.stanbol.enhancer.servicesapi.ChainException: Execution timeout > after 60.001sec (timeout:60sec) for ContentItem > <urn:content-item-sha1-f95a248b47f0a40b2e2f05e97e8b4f6b3a38ee6a> > > Status: > > -EDFImplantationLinking: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusInProgress >> > > -langdetect: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusCompleted >> > > -EDFAcronymeLinking: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusInProgress >> > > -talismane-nlp: < > http://stanbol.apache.org/ontology/enhancer/executionmetadata#StatusCompleted >> -- | Rupert Westenthaler [email protected] | Bodenlehenstraße 11 ++43-699-11108907 | A-5500 Bischofshofen
