db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mike Matrigali <mikem_...@sbcglobal.net>
Subject Re: Derby in large scale deployment ?
Date Fri, 05 Jan 2007 00:15:13 GMT
Also note that derby may not be able to recognize that multiple derby
instances are running if you place the database on a network drive as
it uses per machine JVM locks to catch multiple instances incorrectly
running against the same image.  Not catching this can lead to 
corruptions depending on what the multiple instances are doing.

David Van Couvering wrote:
> I think in general Derby should be able to handle this, but I think you 
> need to put it through it's paces, exercising with the types of 
> operations you'll be submitting to it.  Important things like, are you 
> mostly-read or mostly-write or half and half?  Are you doing mostly 
> single-record primary key operations, or are there lots of joins and sorts?
> In temrs of multiple Derby instances sharing the same Derby database, 
> no, they can not simultaneously.  You can do this if you run Derby as a 
> network server, but not in embedded mode.  In embedded mode, a single 
> instance owns a lock on the database and no other instances are allowed 
> to open connections to it.
> David
> vissu wrote:
>> Hi All,
>> We are contemplating of placing Derby in front of Oracle.  We would 
>> have to
>> distribute 10B rows (in Oracle) across 10 Derby Servers - each would be
>> holding about 1B rows. Rowsize is about 300 bytes.
>> Can Derby handle this volume of data ?. 
>> Secondly, can multiple Derby instances share same Derby Database ?. In 
>> other
>> words, when one Derby updates a rows, will other Derby instances see it -
>> assuming all Derby instances use file based instance and the db 
>> diretcory is
>> located on a network location.  I understand that C-JDBC and other 
>> solutions
>> are fasible, but we dont want to use JGroups.
>> Thanks for any insights and help.
>> Vissu

View raw message