Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Linux & Oracle High Availability
Hi!
Thanks for your comment!
Pete Sharman <psharman_at_us.oracle.com> schrieb in im Newsbeitrag:
38567044.9F0A3C43_at_us.oracle.com...
> 2. Standby database. I don't really understand your comment about that not
being
> nice. Lots of people use it, and with the new functionality in 8i it really
is
> quite suitable for a high availability scenario.
>
> 3. Multi-master replication. The major issue here is (as it always has been)
the
> business decisions to handle conflict. Technically, it works fine.
It seems to me that replication allows me to get more power out of a number of
machines: Assuming that I have two Linux servers running Oracle, in a backup
scenario, one server would just stay in the backup as a fail-over if something
bad happens to the main server. I cannot use the backup server for queries.
In a replication scenario, it seems to me that I can use the power of both
servers. I can query both servers. As far as I understood the backup solution, I
cannot query the backup server!
Thus in replication, I have two servers working. If one fails, the other one
will have to do the whole work, but when I start-up the failed server again, it
will automatically (replication!) set itself to the current state of data with
the help of the still performing other server.
> Which of these three suits you is going to be site and even application
dependent,
> and you haven't provided enough information for us to advise on this. One
comment I
> would make though, is why do you want a highly available development
environment,
> unless it's to test a highly available Production environment?
Sorry, you're of course right. I was not very exact in my writing. Of course the
whole thing is intended to get into production at some point in time...
I'm just planning to set up the different scenarios in a development environment to perform some tests concerning performance. And as I wrote, from my point of view it seems to me that a replication scenario will give me the most concerning performance...
Once again, thanks for your comments!
Bye,
Gerald
Received on Tue Dec 14 1999 - 15:07:51 CST