search

chenejac / VIVOTestMigrationJIRAClosed

0 stars 0 forks source link

VIVO-427: ABox is recomputed for no apparent reason #953

Closed chenejac closed 6 years ago

chenejac commented 11 years ago

Jim Blake (Migrated from VIVO-427) said:

I took some Indiana data which was at release 1.5 and loaded into a 1.6 (rc1) system. It took several hours to migrate and re-index. I did timing tests, and stopped Tomcat.

The next time I started Tomcat with the same code and against the same model, another full re-inference began. See attached log.

Although the log suggests that the re-inference completed, the process was still fully using the CPU on my dev-machine after 2 days had passes.

chenejac commented 11 years ago

Jim Blake said:

Brian - I killed Tomcat and restarted it. VIVO initialized normally.

I will try to reproduce this over the weekend, starting back at 1.5 data and migrating, waiting for quiet, and then restarting. If I can't reproduce the problem, I'll close this issue.

chenejac commented 11 years ago

Brian Lowe said:

It recomputed because it detected a change in the geopolitical abox. I had the same thing happen to me exactly once and then never saw it again. No idea what could account for it.

2013-10-23 15:36:11,673 INFO [FileGraphSetup] Updating /users/jeb228/LoadTesting/versions/indiana1.6/data/rdf/abox/filegraph/geopolitical.abox.ver1.1-11-18-11.owl because graphs are not isomorphic 2013-10-23 15:36:13,640 INFO [FileGraphSetup] dbModel: 22317 ; fileModel: 22317

Note that the old and new models are the same size. The comparison uses Jena's isIsomorphicWith() method to compare the models, since they contain blank nodes.

chenejac commented 11 years ago

Jim Blake said:

I tried re-doing this over the weekend. However, the process aborted after about 8 hours, with this message 2013-10-27 02:52:51,632 ERROR [ABoxRecomputer] Exception while reconciling the current and recomputed ABox inference model for class subsumption inferences. Halting processing. com.hp.hpl.jena.sdb.sql.SDBExceptionSQL: SQLException in executing SQL statement at com.hp.hpl.jena.sdb.compiler.SDB_QC.exec(SDB_QC.java:72) ... Caused by: com.mysql.jdbc.exceptions.jdbc4.CommunicationsException: Communications link failure ... Caused by: java.io.EOFException: Can not read response from server. Expected to read 144 bytes, read 59 bytes before connection was unexpectedly lost.

Full logs are attached.

I will go back to the 1.5 data and restart the migration.

chenejac commented 11 years ago

Jim Blake said:

I restarted from the 1.5 data. The migration went without incident. On subsequent restarts of Tomcat, the ABox was not re-inferenced.