On Fri, 1 Dec 2000, Gordan Bobic wrote:
> > The proxy monitors and forwards the requests from the clients to the
> > database servers. If it is a read-only request the query is forwarded to
> > the databaseserver currently experiencing the lowest load/most free
> > memory, otherwise it is sent to all database servers.
>
> Surely, you have to send off a query to all servers that the specific
> tables you are searching are split across. How are you handling splitting
> of the data?
I'm not ;) All the database servers of the clusters simply each contain
a copy of all the data. I know this isn't an optimal solution, but this
project is only running for a very short period (around 2 months)
> > read-only queries, such as search engines and so on.
>
> This implies lots of servers with identical data on them. Am I right?
>
> Could you tell us a bit more about your project (if it's not breaching any
> non-disclosure agreements, that is)? It could be a good starting point for
> what we are talking about implementing.
There is no nda's involved in this project. We are planning on releasing
the software under GPL after the deadline (22th of December).
There isn't much more to tell about the system at this time. Currently we
are finishing off the lower levels of the systems like database driver
(currently only postgres), status daemons through snmp and so on. At this
time the actual scheduling hasn't been implemented.
--
Bye, Peter Korsgaard