Re: Replication in databases

From: Christoph Rupp <cruppstahl_at_gmail.com>
Date: Fri, 17 Oct 2008 05:44:25 -0700 (PDT)
Message-ID: <c3bab2d0-ccbe-41d3-8605-a57acf0466a6_at_q35g2000hsg.googlegroups.com>


On Oct 17, 8:21 am, David BL <davi..._at_iinet.net.au> wrote:
> On Oct 17, 5:18 am, Christoph Rupp <cruppst..._at_gmail.com> wrote:
>
> > I'm implementing my own database, and therefore i was thinking about
> > the same question. (i have not yet implemented recovery, because
> > there's so much other stuff to do).
>
> > Currently my idea is to send the log over the net, because in my
> > design i do not log physical pages, instead i just log the modified
> > key/record items.
>
> I recall your post 6 months ago :)
>
> Are your log records idempotent?

Yes, that was me :)

however, 6 months ago i implemented logging for my first database - hamsterdb (hamsterdb.com). It's physical logging, writing modified database pages. hamsterdb is good for embedded devices, but has no concurrency and really bad support for transactions.

3 months ago i started writing hamsterdb2, which has a high level of concurrency, full support for transactions and record-level logging instead of page-level logging. and this is still work in progress.

but you can be sure that sooner or later i will continue asking on comp.databases.theory strange questions about different transaction isolation levels and other esoteric stuff :) Received on Fri Oct 17 2008 - 14:44:25 CEST

Original text of this message