Version: unspecified (using KDE 4.5.95) OS: Linux If virtuoso is killed while nepomuk indexes the indexing does not stop and nepomuk spams the .xsession-errors file with [/usr/bin/nepomukservicestub] "/usr/bin/nepomukservicestub(20818)" Soprano: "SQLExecDirect failed on query 'sparql insert into graph ... Expected beahviour: Stop indexing if there is no db available. Since this leads to .xsession-errors increasing in size at ~1MB per second it may as well fill you home and prevent you from logging in or even crash applications that would need some space in home to save files etc. Reproducible: Always
Apart from the fact that this needs fixing indeed: how did Virtuoso die? Did you kill it or did it crash?
I had to kill it because it was using one core all the time.
Git commit 7657a1f7f3b9ee352ac25b2ca4802ca5d654f7d0 by Sebastian Trueg. Committed on 25/10/2011 at 10:27. Pushed by trueg into branch 'KDE/4.7'. Re-create the repository in case of an un-scheduled shutdown of the DB. This is a hacky way of ensuring that Nepomuk even works after a crash or a third-party kill of the Virtuoso instance. "Hacky" because it does not care about clients which will get some query errors during the restart. BUG: 263730 FIXED-IN: 4.7.3 M +21 -2 nepomuk/services/storage/repository.cpp M +1 -0 nepomuk/services/storage/repository.h http://commits.kde.org/kde-runtime/7657a1f7f3b9ee352ac25b2ca4802ca5d654f7d0
Git commit cc5cf57c7e5be45fbf757414994290dd58ea1231 by Sebastian Trueg. Committed on 25/10/2011 at 10:27. Pushed by trueg into branch 'master'. Re-create the repository in case of an un-scheduled shutdown of the DB. This is a hacky way of ensuring that Nepomuk even works after a crash or a third-party kill of the Virtuoso instance. "Hacky" because it does not care about clients which will get some query errors during the restart. BUG: 263730 FIXED-IN: 4.7.3 M +21 -2 nepomuk/services/storage/repository.cpp M +1 -0 nepomuk/services/storage/repository.h http://commits.kde.org/kde-runtime/cc5cf57c7e5be45fbf757414994290dd58ea1231
Git commit ceaa1e22a5279bdef0f451d09c0d27bbaa447035 by Sebastian Trueg. Committed on 25/10/2011 at 10:20. Pushed by trueg into branch 'master'. Re-create the repository in case of an un-scheduled shutdown of the DB. This is a hacky way of ensuring that Nepomuk even works after a crash or a third-party kill of the Virtuoso instance. "Hacky" because it does not care about clients which will get some query errors during the restart. BUG: 263730 M +21 -2 services/storage/repository.cpp M +1 -0 services/storage/repository.h http://commits.kde.org/nepomuk-core/ceaa1e22a5279bdef0f451d09c0d27bbaa447035