jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sven Seibold" <sseib...@translations.com>
Subject Re: Any way to fix broken node
Date Wed, 01 Oct 2008 15:21:41 GMT

Hi Stefan,

thanks for the info. I'll take a look and see if it at least fixes the 
corruption somehow. We already created new nodes for the data, but 
getting the corrupt nodes out will surely not hurt!

I doubt there is a deployment issue. I am sure nothing else is working 
on that repository/workspace/DB . There is only a single instance and 
possibly multiple sessions.
My gut feeling says this is caused by some OutOfMemory problems with the 
application. It happened the first time in 2+ years with 13000+ of these 
nodes in total. All broken 11 nodes were created and updated at about 
the same time. We likely never will get a final answer :(


Stefan Guggisberg wrote:
> hi sven
> On Tue, Sep 30, 2008 at 1:26 PM, Sven Seibold <sseibold@translations.com> wrote:
>> Hi all,
>> we are using Jackrabbit 1.4.5 with the MSSqlPersistenceManager . For a still
>> unknown reason some nodes seem to have been broken and are in a corrupt
>> state.
>> - node.getNodes() does not list these nodes anymore at all
>> - node.getNode(path) gives an PathNotFoundException
>> - node.addChild(path) at the same time gives a ItemExistsException
>> - trying to access an apparently existing child node of these gives an error
>> like
>> javax.jcr.RepositoryException: failed to retrieve state of intermediary
>> node: 1891ab0d-55d5-4947-8ad9-4063e8b3145f
>> I have tried various things but cannot seem to fix these nodes. Any try to
>> access these nodes in order to do something with them ends up with a
>> PathNotFoundException while any try to create these nodes ends with an
>> ItemExistsException .
>> Is there any way to restore recover these nodes (no versioning is used)?
> while it's probably not possible to recover the 'lost' nodes you should be
> able to fix the inconsistency, see
> http://www.nabble.com/Need-Help-%21-tt18376185.html#a18440018.
> i'd really like to know what lead to this inconsistency. usually it's
> caused by incorrect
> deployment (e.g. multiple repository instances accessing the same db db schema).
> cheers
> stefan
>> Thanks in advance for any help!
>> Cheers,
>>  Sven

Sven Seibold <sseibold@translations.com>
Senior Software Developer
Translations.com <http://www.translations.com/>

View raw message