http://blog.gmane.org/gmane.comp.apache.uima.devel hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.comp.apache.uima.devel/5872 UIMA AS Service Should Expose Cas Pool Size for Monitioring via JMX ------------------------------------------------------------------- Key: UIMA-1241 URL: https://issues.apache.org/jira/browse/UIMA-1241 Project: UIMA Issue Type: New Feature Components: Async Scaleout Reporter: Jerry Cwiklik The internal Cas Pool of the UIMA AS service needs to be exposed via JMX for monitoring. Currently, only the Cas Multiplier's Cas Pool is exposed. All Cas Pools are JMX enabled already. Modifiy JmxMonitor to query JMX Server for the Cas Pool size. Jerry Cwiklik (JIRA 2008-12-01T16:27:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5869 I can't print page 32 of the developer's guide. This is on Windows with both Acrobat 8 and 9. Has anybody else seen this? --Thilo Thilo Goetz 2008-12-01T12:43:39 http://comments.gmane.org/gmane.comp.apache.uima.devel/5868 Deployment Descriptor Editor resets CAS Pool size for aggregates when you change number of instances of a delegate ------------------------------------------------------------------------------------------------------------------ Key: UIMA-1240 URL: https://issues.apache.org/jira/browse/UIMA-1240 Project: UIMA Issue Type: Bug Components: Async Scaleout Affects Versions: 2.3AS Reporter: Adam Lally Assignee: Tong Fin I have an async aggregate with multiple delegates. When I set the number of instances of one of the delegates (it happened to be a CAS Multiplier, but I'm not sure that matters), the DDE automatically reset the CAS Pool size of the entire aggregate to the same value! It took me quite a while to figure out that this was the reason my application started running more slowly. Adam Lally (JIRA 2008-11-26T22:36:46 http://comments.gmane.org/gmane.comp.apache.uima.devel/5861 Fix "Consumer Closed" Exception Thrown By ActiveMQ --------------------------------------------------- Key: UIMA-1239 URL: https://issues.apache.org/jira/browse/UIMA-1239 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik ActiveMQ seems to close a connection to a broker if the connection is not actively used. This results in different types of exceptions thrown by a listener on an input queue. Sometimes the exception is " Broken Pipe", in other times "Consumer Closed". This error prevents the service from receiving messages and leads to a hang as CASes sent by a client never come back Jerry Cwiklik (JIRA 2008-11-26T19:00:45 http://comments.gmane.org/gmane.comp.apache.uima.devel/5853 Fix NullPointerException When Sending Reply to Uima AS Client ------------------------------------------------------------- Key: UIMA-1238 URL: https://issues.apache.org/jira/browse/UIMA-1238 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik There is a NPE when sending a reply back to a client from Uima AS aggregate. Caused by: java.lang.NullPointerException at java.util.concurrent.ConcurrentHashMap.put(ConcurrentHashMap.java:846) at org.apache.uima.aae.spi.transport.vm.UimaVmMessage.addStringProperty(UimaVmMessage.java:52) at org.apache.uima.aae.controller.AggregateAnalysisEngineController_impl.sendVMMessage(AggregateAnalysisEngineController_impl.java:1594) at org.apache.uima.aae.controller.AggregateAnalysisEngineController_impl.replyToClient(AggregateAnalysisEngineController_impl.java:1552) at org.apache.uima.aae.controller.AggregateAnalysisEngineController_impl.finalStep(AggregateAnalysi Jerry Cwiklik (JIRA 2008-11-25T17:57:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5851 Fix "Cas Not Found In CasManager Cache" Error in the Parallel Step ------------------------------------------------------------------ Key: UIMA-1237 URL: https://issues.apache.org/jira/browse/UIMA-1237 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik There is an exception when processing a secondary reply from a delegate in a parallel step. The exception indicates that the CAS id received in a reply message from one of the delegates in the parallel step does not exist in the cache. The exception happens in the testcase that tests disable of one of the delegates in the parallel step. Jerry Cwiklik (JIRA 2008-11-25T16:57:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5849 Fix a hang in the testProcessParallelFlowWithDelegateDisable testcase ---------------------------------------------------------------------- Key: UIMA-1236 URL: https://issues.apache.org/jira/browse/UIMA-1236 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik This testcase occasionally hangs after the delegate is disabled and its reply queue listener killed. Jerry Cwiklik (JIRA 2008-11-25T16:15:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5829 In preparation for the release, please take a look at http://incubator.apache.org/uima/doc-uimacpp-huh.html Thanks, Eddie Eddie Epstein 2008-11-23T19:59:54 http://comments.gmane.org/gmane.comp.apache.uima.devel/5828 Changes for the first UIMACPP release. -------------------------------------- Key: UIMA-1235 URL: https://issues.apache.org/jira/browse/UIMA-1235 Project: UIMA Issue Type: New Feature Components: Website Reporter: Eddie Epstein Assignee: Eddie Epstein Download links, getting started guide, etc. Eddie Epstein (JIRA 2008-11-23T17:27:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5817 Cleanup Testcases to Handle ServiceShutdownException Quietly ------------------------------------------------------------ Key: UIMA-1234 URL: https://issues.apache.org/jira/browse/UIMA-1234 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik UIMA AS testcases dump ServiceShutdownException to stdout. Modify testcase code to ignore this exception. Service sends a ShutdownException to its client when it terminates. Jerry Cwiklik (JIRA 2008-11-21T21:45:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5814 Log Outgoing Message Sizes Destined for Remote Queue ---------------------------------------------------- Key: UIMA-1233 URL: https://issues.apache.org/jira/browse/UIMA-1233 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik Log the size of CAS before sending it to a remote queue. Jerry Cwiklik (JIRA 2008-11-21T20:21:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5813 Fix Synchronization Problem in the UIMA AS code ----------------------------------------------- Key: UIMA-1232 URL: https://issues.apache.org/jira/browse/UIMA-1232 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik UIMA AS Aggregate code synchronizes code that sends a reply (CAS) back to a client. Part of sending the CAS back to the remote client is serialization of the CAS. This takes time. While the CAS is being serialized, other threads are blocked causing a bottleneck. Move the code that sends a reply outside of the synchronized block. Jerry Cwiklik (JIRA 2008-11-21T20:15:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5812 UIMA AS Doesnt Expose to JMX Sizes of Java Queues Used For Internal Messaging ----------------------------------------------------------------------------- Key: UIMA-1231 URL: https://issues.apache.org/jira/browse/UIMA-1231 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik Uima AS uses java queues for internal messaging. The depths of these queues are not visible in jmx. Add JMX support to both input and output queues. Modify JMX Monitor to show the depth of the java queues. Jerry Cwiklik (JIRA 2008-11-21T20:11:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5798 When parsing an aggregate descriptor, should parse a shared type system file only once -------------------------------------------------------------------------------------- Key: UIMA-1230 URL: https://issues.apache.org/jira/browse/UIMA-1230 Project: UIMA Issue Type: Improvement Components: Core Java Framework Affects Versions: 2.2.2 Reporter: Adam Lally Assignee: Adam Lally Priority: Minor Fix For: 2.3 Frequently an aggregate will contain several primitives that all import a single shared type system descriptor. Currently, the shared type system descriptor will be separately parsed for each of the primitives, and each primitive descriptor will contain a separate copy of the TypeSystemDescription object. When there are a large number of primitives and a large number of types in the shared type system, this parsing takes a long time and the resulting TypeSystemDescription objects consum Adam Lally (JIRA 2008-11-20T19:42:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5796 Hi guys, Not sure if I am supposed to initiate the vote myself as I am not a committer. Are you in favour of adding the Tika Annotator to the sandbox? Best, Julien Julien Nioche 2008-11-20T14:16:51 http://comments.gmane.org/gmane.comp.apache.uima.devel/5793 Hi everyone, The Apache UIMA PPMC is very pleased to announce that Tong Fin and Jerry Cwiklik have accepted our invitation to become Apache UIMA committers. They have been voted in by the Incubator PMC, following the normal procedures for this. Tong has contributed the CAS Viewer and is working with Jörn on integrating that work with the CAS Editor. Tong has extensive experience in developing Eclipse plugins, and has been helpful to others in this area. Jerry has been contributing to UIMA since its inception, and recently has been the main designer and implementor of UIMA-AS - the Asynchronous messaging Scaleout that allows for flexibly scaling out of parts of the UIMA pipeline. Jerry was also the principle developer of the Collection Processing Engine design and implementation in the core UIMA framework. We're thrilled to have both of them join us in our ongoing UIMA work, and look forward to their continued contributions. Congratulations, and welcome aboard, Tong & Jerry! -Marshall Marshall Schor 2008-11-19T22:34:36 http://comments.gmane.org/gmane.comp.apache.uima.devel/5790 Pears are UIMA analysis components that are wrapped/packaged to include the classpath & data path needed to run them. A Pear can be run as a component in an Aggregate. When it is run, the current Java Classpath and data path are switched from their current values to the values specified in the Pear, while the Annotator is running, and then switched back. One of the things that the Pear specification allows is the setting of "environment variables". This capability was initially put in for legacy applications, many of which were written in non-Java languages, to allow them to run unmodified, if they depended on some environmental variables to be set. The initial "runner" for these kinds of Pears was thought to be special repository code, which might use a web interface to allow users to specify analysis components to run against some data. UIMA itself did not specify how these environmental variable settings were to be used. The current Pear Wrapper implementation which runs Pears within a UIMA pipelin Marshall Schor 2008-11-17T20:58:01 http://comments.gmane.org/gmane.comp.apache.uima.devel/5785 Ken Coar (our mentor) said our status file was out of date. I took a look and found 3 things to update: one was to add Jörn to the committers list, another was to include our private mailing list in the list of mailing lists, and the third was to point to where all our status reports were (on the incubator wiki). I think I got everything updated ok (may have to wait a bit for replication). -Marshall Marshall Schor 2008-11-14T19:43:05 http://comments.gmane.org/gmane.comp.apache.uima.devel/5783 Add Private (not shared) Cache to UIMA AS Controllers For Storing Intermediate State for Each CAS ------------------------------------------------------------------------------------------------- Key: UIMA-1229 URL: https://issues.apache.org/jira/browse/UIMA-1229 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Jerry Cwiklik UIMA AS uses one (shared) instance of a cache where intermediate state information is kept (and changed) for each CAS during processing. This cache is shared among all colocated components. In complex deployments involving colocated, multi-level aggregates CAS states changed by one aggregate cause problems in other aggregates. Particularly when making decisions when to release a CAS. In such deployments hangs may occur. Add a private (non-shared) cache to each UIMA AS controller where the state of each CAS is kept. CAS instance should be kept in the global InProcessCache that is shared b Jerry Cwiklik (JIRA 2008-11-12T20:39:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5769 Merging results from a parallel step fails when a feature is an array of pre-existing annotations ------------------------------------------------------------------------------------------------- Key: UIMA-1228 URL: https://issues.apache.org/jira/browse/UIMA-1228 Project: UIMA Issue Type: Bug Components: Async Scaleout Reporter: Burn Lewis Fix For: 2.3AS Found when doing 2.2.2-style merging (non deltaCas) of new arrays referencing preexisting annotations, but also fails with delta CASes. Burn Lewis (JIRA 2008-11-10T16:15:44 http://comments.gmane.org/gmane.comp.apache.uima.devel/5768 Help users debug binary CAS serialization problems -------------------------------------------------- Key: UIMA-1227 URL: https://issues.apache.org/jira/browse/UIMA-1227 Project: UIMA Issue Type: Improvement Components: Async Scaleout Reporter: Eddie Epstein In order to use binary CAS serialization for remote services, both client and service must have identical type systems. If not the same, binary deserialization is likely to throw exceptions. The service request handler should catch these exceptions when using binary deserialization and add message suggesting a likely cause of the problem: unequal type systems. Eddie Epstein (JIRA 2008-11-07T17:06:45 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.apache.uima.devel