Re: beagle doesn't remove stale locks on startup

[Alec Burney <madbull gmail com>]

On Mon, 2006-06-19 at 08:19 -0700, Debajyoti Bera wrote: 
> > beagle seems not not remove its locks on the next startup after a very
> > unclean shutdown, meaning that once it tries to gain a lock and fails
> > because of the existing lock, it will stop indexing alltogether for that
> > session. then it will not shutdown cleanly or remove the existing lock,
> > meaning beagle is basically useless until you remove the lock by hand.
> >
> > before:
> >
> > Warn: Exception caught while executing
> > Beagle.Daemon.ConnectionHandler:Void HandleConnection()
> > Warn: System.IO.IOException: Lock obain timed
> > out:
> > /home/madbull/.beagle/Indexes/FileSystemIndex/Locks/lucene-40acc3cb58f13242
> >ebb56fd9feecc673-write.lock -- pid  -- process exists in [0x0010e] (at
> > /root/beagle/beagled/Lucene.Net/Store/Lock.cs:91)
> > Lucene.Net.Store.Lock:Obtain (Int64 lockWaitTimeout)
> > in [0x00039]
> > (at /root/beagle/beagled/Lucene.Net/Index/IndexReader.cs:547)
> > Lucene.Net.Index.IndexReader:AquireWriteLock ()
> > in [0x00014]
> > (at /root/beagle/beagled/Lucene.Net/Index/IndexReader.cs:576)
> > Lucene.Net.Index.IndexReader:Delete (Int32 docNum)
> > in [0x0001f]
> > (at /root/beagle/beagled/Lucene.Net/Index/IndexReader.cs:605)
> > Lucene.Net.Index.IndexReader:Delete (Lucene.Net.Index.Term term)
> > in [0x000a4] (at /root/beagle/beagled/LuceneIndexingDriver.cs:126)
> > Beagle.Daemon.LuceneIndexingDriver:Flush_Unlocked
> > (Beagle.Daemon.IndexerRequest request)
> > in [0x0000f] (at /root/beagle/beagled/LuceneIndexingDriver.cs:90)
> > Beagle.Daemon.LuceneIndexingDriver:Flush (Beagle.Daemon.IndexerRequest
> > request)
> > in [0x00079]
> > (at /root/beagle/beagled/IndexHelper/RemoteIndexerExecutor.cs:69)
> > Beagle.IndexHelper.RemoteIndexerExecutor:Execute (Beagle.RequestMessage
> > raw_request)
> > in [0x00232] (at /root/beagle/beagled/Server.cs:275)
> > Beagle.Daemon.ConnectionHandler:HandleConnection ()
> > in (wrapper delegate-invoke) System.MulticastDelegate:invoke_void ()
> > in [0x0005a] (at /root/beagle/Util/ExceptionHandlingThread.cs:54)
> > Beagle.Util.ExceptionHandlingThread:ThreadStarted ()
> >
> >
> > # ls .beagle/Indexes/FilesystemIndex/Locks
> >  lucene-0b58ff34d9f20b2c4e4d5f10dfe8bb79-write.lock
> >  lucene-40acc3cb58f13242ebb56fd9feecc673-write.lock
> >
> > so, remove them:
> > # rm .beagle/Indexes/FilesystemIndex/Locks/*lock
> >
> > and we come to the after:
> > no problem at all. beagle runs as normal again.
> > i would submit a patch, but i am by no means a C# programmer.
> 
> A few weeks back a patch was checked in CVS to fix this behaviour. Please give 
> that a try (or wait for the next release, which is around the corner). Thanks 
> for reporting.
> 
> > PS: this is the right place to put stuff like this, right? sorry if it's
> > not.
> 
> This is indeed the right place. Welcome.
> 
> - dBera
> 

Thanks, for the responses, guys. I forgot to mention that i'm running on
CVS from about a month ago. I'm the proud owner of a shiny new build,
0.2.7, now.
anyway, i'm glad to hear that this problem is fixed. you've really got a
great indexing engine going.