I didn't pay full attention to this thread, but it sounds like somebody may be interested in RuntimeShutdownHook (or some similar name) as a place to try to release the locks.
Otis --- Joseph Ottinger <[EMAIL PROTECTED]> wrote: > On Tue, 11 Jan 2005, Doug Cutting wrote: > > > Joseph Ottinger wrote: > > > As one for whom the question's come up recently, I'd say that > locks need > > > to be terminated gracefully, instead. I've noticed a number of > cases where > > > the locks get abandoned in exceptional conditions, which is > almost exactly > > > what you don't want. > > > > The problem is that this is hard to do from Java. A typical > approach is > > to put the process id in the lock file, then, if that process is > dead, > > ignore the lock file. But Java does not let one know process ids. > Java > > 1.4 provides a LockFile mechanism which should mostly solve this, > but > > Lucene 1.4.3 does not yet require Java 1.4 and hence cannot use > that > > feature. Lucene 2.0 is likely to require Java 1.4 and should be > able to > > do a better job of automatically unlocking indexes when processes > die. > > Agreed - but while there are some situations in which releasing locks > is > "difficult" (i.e., JVM catastrophic shutdown), there are others in > which > attempts could be made via finally blocks, etc. > > ----------------------------------------------------------------------- > Joseph B. Ottinger > http://enigmastation.com > IT Consultant > [EMAIL PROTECTED] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
