Corrupted nodes after disk space exceeded

[maneo <ma...@man.poznan.pl>]

Hi,

I have problem with consistency of my repository, recently we've 
experienced problem with hard disk space on server which hosts our 
service. After those problems (service operated with 0 bytes of free 
space for some time) our repository throws exceptions like this below:

Exception while sorting nodes in document order: 
javax.jcr.ItemNotFoundException: failed to build path of 
6ead8aaf-377a-450e-9e82-f43deda5e911: 
ebccc515-7b7a-4334-b660-68a627c22f85 has no child entry for 
6ead8aaf-377a-450e-9e82-f43deda5e911

To fix this I exported (using Session.exportXML) all nodes (created by 
my app) from corrupted repository and placed those nodes in brand new 
repository (using Session.importXML). As of corrupted nodes everything 
seems to be fine now. But this new repository is 3 times smaller than 
original one! I've checked, it's not a matter of Lucene indices, 
difference lies in size of database. Do you have any idea what is the 
reason of such a difference?

Maybe there is a better way to remove corrupted nodes, I cannot remove 
them using standard API (node.remove() - I can fetch them using to UUID) 
but my knowledge is rather limited in jcr area.

I'm using jackrabbit-1.3.1.

Thanks in advance,
Adam

Re: Corrupted nodes after disk space exceeded

[maneo <ma...@man.poznan.pl>]

Sven Schliesing wrote:
> Hi Adam,
> 
> what database are you using? I experienced derby to grow even if you 
> delete content from time to time[1].
> 
> Maybe your old db was just bloated with unused space.
> 

I use derby, so this may be the cause. Thanks for that link :).

thanks. Adam

Re: Corrupted nodes after disk space exceeded

[Sven Schliesing <sc...@subshell.com>]

Hi Adam,

what database are you using? I experienced derby to grow even if you 
delete content from time to time[1].

Maybe your old db was just bloated with unused space.


Greetings,

Sven


[1] http://db.apache.org/derby/docs/10.1/adminguide/cadminspace21579.html