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: Does Derby support Transaction Logging ?
Date Mon, 20 Nov 2006 18:44:27 GMT
Yes, derby is not there yet -- but with the online backup work done in
10.2 it is very close.  The problem is that there is no support for
coordination of taking and applying the logs in step 3.  Also what
do you want to be doing on server_2 during/after this process.

I think this is a smallish project to create a readonly standby backup
server (on the order of month's) for someone
who already knows how to work with derby.  The missing pieces now
o enhance the online backup code to allow for copying the log files
   in some coordinated fashion.  Currently it just keeps the existing
   logs where they are forever.  An easy first step would be to just
   enable a callout everytime there is a log switch.

o Allow for a standby server to sit around, allow only readonly
   access to data.  Apply log files as they are fed in.  Maybe get locks
   on log record application (existing code for this exists as part of
   XA redo).

Now allowing write access on the standby while taking log files is
way more difficult.

Duncan Groenewald wrote:
> Thanks, I just read the material on the transaction logs.  However I 
> still don't see how I would be able to achieved what I want which is:
>   1. Set up a new derby database server on another server (server_2)
>   2. Restore the database from a database backup from server_1
>   3. Copy server_1 logs every 15 minutes and load them into server_2
> Duncan Groenewald wrote:
>> Yes, I want to load the transaction logs into a copy of the database 
>> in a simple form of replication.  Is there an alternative way of doing 
>> this ?
>> Dan Scott wrote:
>>> On 15/11/06, Stanley Bradbury <Stan.Bradbury@gmail.com> wrote:
>>>> Duncan Groenewald wrote:
>>>> > Does derby support transaction logging ? and if so how to dump/access
>>>> > them ?  If not any plan to do so ??
>>>> >
>>>> I don't understand what you are asking about but can tell you that 
>>>> Derby
>>>> maintains a transaction log for rollback and recovery purposes.  It is
>>>> not intended to be accessed directly by and application.  There is some
>>>> access to these records via Triggers.  The architecture information on
>>>> the website contain good descriptions of the transactional capabilities
>>>> of Derby.
>>> A completely wild guess would be, from a MySQL perspective, that
>>> somebody wants to try to implement a form of replication by
>>> periodically capturing the transaction logs from a master instance of
>>> an Apache Derby database and then applying those logs to a replicated
>>> instance of an Apache Derby database.
>>> Dan

View raw message