Re: weblogic cluster, index on NFS and locking problem
David please let me know if you submit your code. So we could take a look at it. It sounds very good for us. Dmitri We work on NFS and have had major problems with locking, which often leads to the indices becoming corrupt. Our solution was to replace file locking with a database system. I can release the code but I'm not sure of the process or where to put it. It basically two classes one that extends the Directory class and one that deals with the database interaction. David Townsend -Original Message- From: Dmitri Ilyin [mailto:[EMAIL PROTECTED] Sent: 04 February 2004 09:49 To: [EMAIL PROTECTED] Subject: Re: weblogic cluster, index on NFS and locking problem What is it good for??? unfortunately i don't have any access to NFS server. It runs at customers in production environment. Suggestion: make sure the NFS lock daemon (lockd) is running on the NFS server. Peter - Original Message - From: "Dmitri Ilyin" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, February 03, 2004 1:09 PM Subject: weblogic cluster, index on NFS and locking problem Hi, We run our application on weblogic cluster. the lucene index service runs on both server in cluster and they both write to one index directory, shared via NFS. We have experenced a problem with commit.lock file, that seems not to be deleted and stayed in the index directory, so we could not start indexing any more becouse lucene could not create/read commit.lock file. I'm not sure what excatly our problem is. It could be NFS problem or it could be our "usage" problem. We are just starting to use lucene and could made something wrong. We use lucene to index and to search documents. Write/read could be concurent. I saw in the list some messages about problems with lock files on NFS file system. But i could not realy understand what the problem is. How can we improve our solution?? What do we have to do excatly to avoid problem with stayed commit.lock file??? thaks for any advise regards Dmitri - 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] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: weblogic cluster, index on NFS and locking problem
The best way to submit contributions is via Bugzilla. For instance, here is the current queue of contributed code, patches, etc.: http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&email1=&emailtype1=substring&emailassigned_to1=1&email2=&emailtype2=substring&emailreporter2=1&bugidtype=include&bug_id=&changedin=&votes=&chfieldfrom=&chfieldto=Now&chfieldvalue=&product=Lucene&short_desc=%5BPATCH%5D&short_desc_type=allwordssubstr&long_desc=&long_desc_type=allwordssubstr&bug_file_loc=&bug_file_loc_type=allwordssubstr&keywords=&keywords_type=anywords&field0-0-0=noop&type0-0-0=noop&value0-0-0=&cmdtype=doit&order=%27Importance%27 The process consists of: - opening a Bugzilla account - Opening a 'bug' in Bugzilla with the Subject that starts with "[PATCH]" - Attaching the contribution (as ZIP of patches or as diff -u style patch, if it's a smaller patch) Thanks, Otis --- David Townsend <[EMAIL PROTECTED]> wrote: > We work on NFS and have had major problems with locking, which often > leads to the indices becoming corrupt. Our solution was to replace > file locking with a database system. I can release the code but I'm > not sure of the process or where to put it. It basically two classes > one that extends the Directory class and one that deals with the > database interaction. > > David Townsend > > -Original Message- > From: Dmitri Ilyin [mailto:[EMAIL PROTECTED] > Sent: 04 February 2004 09:49 > To: [EMAIL PROTECTED] > Subject: Re: weblogic cluster, index on NFS and locking problem > > > What is it good for??? > unfortunately i don't have any access to NFS server. It runs at > customers in production environment. > > Suggestion: make sure the NFS lock daemon (lockd) is running on the > NFS > > server. > > > > Peter > > > > - Original Message - > > From: "Dmitri Ilyin" <[EMAIL PROTECTED]> > > To: <[EMAIL PROTECTED]> > > Sent: Tuesday, February 03, 2004 1:09 PM > > Subject: weblogic cluster, index on NFS and locking problem > > > > > > > >>Hi, > >> > >>We run our application on weblogic cluster. the lucene index > service > >>runs on both server in cluster and they both write to one index > >>directory, shared via NFS. We have experenced a problem with > commit.lock > >>file, that seems not to be deleted and stayed in the index > directory, so > >>we could not start indexing any more becouse lucene could not > >>create/read commit.lock file. > >> > >>I'm not sure what excatly our problem is. It could be NFS problem > or it > >>could be our "usage" problem. We are just starting to use lucene > and > >>could made something wrong. > >> > >>We use lucene to index and to search documents. Write/read could be > >>concurent. > >> > >>I saw in the list some messages about problems with lock files on > NFS > >>file system. But i could not realy understand what the problem is. > >> > >>How can we improve our solution?? What do we have to do excatly to > avoid > >>problem with stayed commit.lock file??? > >> > >>thaks for any advise > >> > >>regards > >> > >>Dmitri > >> > >> > >> > >>- > >>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] > > > - > 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]
RE: weblogic cluster, index on NFS and locking problem
We work on NFS and have had major problems with locking, which often leads to the indices becoming corrupt. Our solution was to replace file locking with a database system. I can release the code but I'm not sure of the process or where to put it. It basically two classes one that extends the Directory class and one that deals with the database interaction. David Townsend -Original Message- From: Dmitri Ilyin [mailto:[EMAIL PROTECTED] Sent: 04 February 2004 09:49 To: [EMAIL PROTECTED] Subject: Re: weblogic cluster, index on NFS and locking problem What is it good for??? unfortunately i don't have any access to NFS server. It runs at customers in production environment. > Suggestion: make sure the NFS lock daemon (lockd) is running on the NFS > server. > > Peter > > - Original Message - > From: "Dmitri Ilyin" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Tuesday, February 03, 2004 1:09 PM > Subject: weblogic cluster, index on NFS and locking problem > > > >>Hi, >> >>We run our application on weblogic cluster. the lucene index service >>runs on both server in cluster and they both write to one index >>directory, shared via NFS. We have experenced a problem with commit.lock >>file, that seems not to be deleted and stayed in the index directory, so >>we could not start indexing any more becouse lucene could not >>create/read commit.lock file. >> >>I'm not sure what excatly our problem is. It could be NFS problem or it >>could be our "usage" problem. We are just starting to use lucene and >>could made something wrong. >> >>We use lucene to index and to search documents. Write/read could be >>concurent. >> >>I saw in the list some messages about problems with lock files on NFS >>file system. But i could not realy understand what the problem is. >> >>How can we improve our solution?? What do we have to do excatly to avoid >>problem with stayed commit.lock file??? >> >>thaks for any advise >> >>regards >> >>Dmitri >> >> >> >>- >>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] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: weblogic cluster, index on NFS and locking problem
What is it good for??? unfortunately i don't have any access to NFS server. It runs at customers in production environment. Suggestion: make sure the NFS lock daemon (lockd) is running on the NFS server. Peter - Original Message - From: "Dmitri Ilyin" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, February 03, 2004 1:09 PM Subject: weblogic cluster, index on NFS and locking problem Hi, We run our application on weblogic cluster. the lucene index service runs on both server in cluster and they both write to one index directory, shared via NFS. We have experenced a problem with commit.lock file, that seems not to be deleted and stayed in the index directory, so we could not start indexing any more becouse lucene could not create/read commit.lock file. I'm not sure what excatly our problem is. It could be NFS problem or it could be our "usage" problem. We are just starting to use lucene and could made something wrong. We use lucene to index and to search documents. Write/read could be concurent. I saw in the list some messages about problems with lock files on NFS file system. But i could not realy understand what the problem is. How can we improve our solution?? What do we have to do excatly to avoid problem with stayed commit.lock file??? thaks for any advise regards Dmitri - 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]
Re: weblogic cluster, index on NFS and locking problem
Suggestion: make sure the NFS lock daemon (lockd) is running on the NFS server. Peter - Original Message - From: "Dmitri Ilyin" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, February 03, 2004 1:09 PM Subject: weblogic cluster, index on NFS and locking problem > Hi, > > We run our application on weblogic cluster. the lucene index service > runs on both server in cluster and they both write to one index > directory, shared via NFS. We have experenced a problem with commit.lock > file, that seems not to be deleted and stayed in the index directory, so > we could not start indexing any more becouse lucene could not > create/read commit.lock file. > > I'm not sure what excatly our problem is. It could be NFS problem or it > could be our "usage" problem. We are just starting to use lucene and > could made something wrong. > > We use lucene to index and to search documents. Write/read could be > concurent. > > I saw in the list some messages about problems with lock files on NFS > file system. But i could not realy understand what the problem is. > > How can we improve our solution?? What do we have to do excatly to avoid > problem with stayed commit.lock file??? > > thaks for any advise > > regards > > Dmitri > > > > - > 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]