Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Parallel Server & Standby Database

Re: Parallel Server & Standby Database

From: Eugene Firyago <efiryago_at_bisys.com>
Date: 2000/04/26
Message-ID: <8e6upk$jot$1@bob.news.rcn.net>#1/1

Roger is right:
Clustering is a purely _local_ failover solution as all nodes are local to the one database disk structure and are in persistent on-line with the last one. So it protects against a local node failure (not even disk crash) so it does not protect against disaster (fire/earthquake/sabotage). Whereas the standby is "remote" and "off-line" disaster recovery solution and then applicable (at least in theory) to the OPS as well.

Eugene.

"Billy Verreynne" <vslabs_at_onwe.co.za> wrote in message news:3906cb0e.1141433009_at_news.saix.net...
> Mario Bucsics <mario.bucsicsNOmaSPAM_at_siemens.at.invalid> wrote:
>
> >Is anybody out there who knows if it´s possible to create a
> >Standby Database for an Oracle Parallel Server ? I think that
> >this does not work but i´m gratful for any other information.
>
> I fail to undestand why you would want a standby database for OPS. OPS
> is after all a redundant configuration. If one or two nodes go down in
> a three node OPS cluster, you still have one node up and running.
> Short of a catastrophic failure (explosion, fire etc.), it is
> extremely unlikely that you will loose all your nodes in a cluster at
> the same time, thus requiring something like a standby database.
>
> If for whatever reason, you feel that the cluster is not redundant or
> secure enough, rather spend the time and money fixing that, than to
> invest it in a standby database.
>
> You can even configure a cluster for cell redundancy (where each cell
> is a physical set of nodes and disks). For example, you can mirror
> data across cells to ensure that a complete set of database disks are
> up and running, even if a whole cell in the cluster goes belly up.
>
> Mario, I see that you have Siemens in your e-mail address. The cluster
> with cell redudancy I am refering to was in fact a Siemens Nixdorf
> Reliant RM1000 (Pyramid Nile) system. Amazing technology at the time.
>
> A standby databse defeats IMO the whole concept of using massive
> parallel processing/clustering acrhitecure and Oracle Parallel Server.
>
> regards,
> Billy
>
>
Received on Wed Apr 26 2000 - 00:00:00 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US