Problem during documents copy

[freedev]

Hi All,

I had created a new user and I would like to copy all his document into Alfresco, in the user home.
During the copy alfresco hangup (I was transferring about 200 MB) and after few minutes I had to kill tomcat.
Now the problem is that the user home is empty but I see into the directory alf_data/contentstore/ now I have two new directories:

2007/9/28/15/51
2007/9/28/15/52

and I see there is a lot of stuff inside:

$ df -h .
45M     ./51
78M     ./52

The question is, how can I rescue that files?
Or, anyway how I can be sure of the Virtual File System Integrity?

Bye
Vincenzo

[heislord5]

yes I've had a similar problem..even without killing alfresco.  When I'm moving a lot of data and then deleting those folders, I appear to be getting orphaned folders and content often.

[freedev]

Have you found how to recover your data? I had to reinstall everything to get back to a clean status.
And anyway I can not be sure of file system integrity at all.
I think we need (to write ?) some kind of utility capable to check the virtual file system integrity and now I'm investigating this way.

[heislord5]

I don't know any way to fix these types of problems.  I just dropped & recreated the database schema 'alfresco' and deleted the content.

No need to reinstalled the whole application…just wiped out the data and content.

I wish there were a way to fix problems.  Is there any equivalent to Documentum DQL and API to fix problems like this?

Are there integrity checks like the jobs in Documentum to check, report on & fix integrity problems?

[freedev]

I did the same: shutdown the application, drop database and content directory and start the application again.

I have NOT found any integrity check utility.

I think is really important have a such kind of service or utility because in alfresco everything is based on the virtual file system and I could not imagine a operating system like linux that is based on a buggy file system.

[kevinr]

I suggest you rebuild your Lucene indexes. There are several posts that tell how you.

Thanks,

Kevin

[freedev]

Hi Kevin,

I already rebuilded the lucene indexes but there are still files (573 files, about 100MB size) into the alf_data directory that I am sure aren't into the virtual file system.

Looking into the alfresco database instance I found a table, AVM_NODES.
AVM_NODES seems contain all the references to the alf_data files. The field CONTENT_URL explicitly point the files into the alf_data directory.

So now I'm written a simple standalone jpa application that recursively check which files are inside AVM_NODES and which not.
This application I think could be considered the first step of a virtual file system "scandisk". In the following days I'll write the results of my application.

In the meanwhile, please could you help me to find some documentation about the structure of alfresco database?

Bye,
Vincenzo