You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cayenne.apache.org by Peter Schröder <Pe...@freenet-ag.de> on 2006/12/20 15:22:04 UTC

jndi fallback crashes

hi,

i am currently using cayenne with jndi resources. i use the cayenne-modeler as a fallback as discribed here:
http://cwiki.apache.org/CAYDOC/using-jndi.html
which works fine till yesterday, where cayenne told me, that it cant find any resources. in the end i deleted and recreated all data-sources in the modeler, which solved the problem.

today same thing happend again and after poking around in the .cayenne folder i found out that there must be a messed up db.xxx file. i deleted all db.xxx files and replaced them with a working copy i had from yesterday. i didnt change anything in cayenne-modeler, so there has to be something with the fallback. is there a known issue about this?

kind regards 
peter

Re: jndi fallback crashes

Posted by Andrus Adamchik <an...@objectstyle.org>.
Yeah, the known issue is HSQLDB lock not being cleaned on abnormal  
app termination of the modeler. Right now the only solution is the  
one you described - close everything and delete all db[1..9]+.* and  
db.lck files. Going forward we need to switch to some other DB like  
derby or just start using standard java preferences API :-/

Andrus


On Dec 20, 2006, at 4:22 PM, Peter Schröder wrote:

> hi,
>
> i am currently using cayenne with jndi resources. i use the cayenne- 
> modeler as a fallback as discribed here:
> http://cwiki.apache.org/CAYDOC/using-jndi.html
> which works fine till yesterday, where cayenne told me, that it  
> cant find any resources. in the end i deleted and recreated all  
> data-sources in the modeler, which solved the problem.
>
> today same thing happend again and after poking around in  
> the .cayenne folder i found out that there must be a messed up  
> db.xxx file. i deleted all db.xxx files and replaced them with a  
> working copy i had from yesterday. i didnt change anything in  
> cayenne-modeler, so there has to be something with the fallback. is  
> there a known issue about this?
>
> kind regards
> peter
>