Opened 16 years ago

Closed 16 years ago

#1711 defect closed invalid (invalid)

PyLucene doesn't close write indexes properly

Reported by: Jean-Paul Calderone Owned by: amir
Priority: high Milestone:
Component: core Keywords:
Cc: Branch:


This causes a lock to remain in the system which prevents the index from being opened by subsequent runs with this traceback:

2006/05/11 04:13 EDT [-] Traceback (most recent call last):
          File "/home/divmod/Projects/Divmod/trunk/Axiom/axiom/", line 1037, in step
            itemHasMore =,, skip=self.suspended)
          File "/home/divmod/Projects/Divmod/trunk/Axiom/axiom/", line 964, in transact  
            result = f(*a, **k)
          File "/home/divmod/Projects/Divmod/trunk/Axiom/axiom/", line 254, in step
            if listener.step():
          File "/home/divmod/Projects/Divmod/trunk/Axiom/axiom/", line 220, in step
            self._doOneWork(workUnit, _ForwardProcessingFailure)
        --- <exception caught here> ---
          File "/home/divmod/Projects/Divmod/trunk/Axiom/axiom/", line 194, in _doOneWork
          File "/home/divmod/Projects/Divmod/trunk/Mantissa/xmantissa/", line 88, in processItem
            self._index = self.openWriteIndex()
          File "/home/divmod/Projects/Divmod/trunk/Mantissa/xmantissa/", line 349, in openWriteIndex
            PyLucene.IndexWriter(fsdir, analyzer, create),
          File "/usr/lib/python2.4/site-packages/", line 2261, in __init__
            newobj = _PyLucene.new_IndexWriter(*args)
        PyLucene.JavaError: Lock obtain timed out: Lock@/tmp/lucene-b5448e9f96c58cc68d1130b92e20136c-write.lock

This also slows down message delivery by several orders of magnitudes. I'm not sure what right API is to clean these up or prevent them from being created in the first place.

Change History (1)

comment:1 Changed 16 years ago by Jean-Paul Calderone

Resolution: invalid
Status: newclosed
Note: See TracTickets for help on using tickets.