Re: Symantec Storage Foundation for Oracle RAC

Tim,

>I'm somewhat familiar with the terms and distinctions, having worked with
OPS/RAC since 1993 (v7.1.3.0.0), HACMP since 1995, ServiceGuard since 1997, and VCS/SFOR on Solaris since 1999.

I thought you probably were, but you could have just come out and said you don't think RAC is a good/cheap HA solution.

>Many have accepted Oracle's marketing hype that RAC is effective (and
cost-effective) HA. In fact, RAC is a scalability feature that includes some HA features out of necessity

Interesting you should say that. I've worked with RAC for a couple of years now and in my opinion (testing as well as production), RAC provides some good HA when set up correctly and everybody understands the limitations, but scales poorly if the app is not written directly for RAC. Especially applications that are write intensive on a few tables.

Of course, just my $.02. YMMV.

Finn

On 10/30/07, Andrew Kerber <andrew.kerber_at_gmail.com> wrote:
>
> Depending on how its implemented, RAC can allow for HA when your computer
> system fails (power supply goes down on one node for example), or any other
> nodal type harware failures such as faulty network cards, memory, etc. RAC
> is also extremely useful for load balancing. Now if your failure is on the
> storage side, then no, RAC isnt particularly useful as an HA solution. RAC
> however can be part of an HA environment.
>
> On 10/29/07, ken_nguyen_at_comcast.net <ken_nguyen_at_comcast.net > wrote:
> >
> > Tim-
> >
> > I'm not sure what you're trying to get at, and don't really see why you
> > wouldn't come right out and say it, but anyway..
> >
> > So, the particular solution that I'm looking at employs a Global
> > Clustering Option. So, whatever happens at the primary site to cause a
> > failure (earthquake, wild fire, etc..) that the company deems as an outage,
> > would kick off the remote site to come online as the primary database.
> >
> > Ken
> >
> >
> > -------------- Original message --------------
> > From: Tim Gorman < tim_at_evdbt.com>
> > Would an earthquake level just part of a data center? Do wildfires only
> > burn the 3rd server in a rack and not the 4th, sparing the storage in the
> > next cabinet over?
> >
> > In what way does RAC (or any clustering solution, such as SFOR) protect
> > against such threats?
> >
> >
> >
> >
> > ken_nguyen_at_comcast.net wrote:
> >
> > Tim,
> >
> > As Finn mentioned, I'm looking at this to provide HA. This is for a
> > location in SoCal, so its prone to earthquakes, and most recently, wild
> > fires.
> >
> > Ken
> >
> >
> > -------------- Original message --------------
> > From: "Finn Jorgensen" <finn.oracledba_at_gmail.com><finn.oracledba_at_gmail.com>
> > Tim,
> >
> > RAC/SFOR/HACMP/VCS etc are for high availability. Not disaster recovery.
> > As such, it's for the type of "disasters" that involve losing 1 server.
> > Anything else you would need a DR setup to handle.
> >
> > Finn
> >
> >
> > On 10/29/07, Tim Gorman <tim_at_evdbt.com> wrote:
> > >
> > > Well......out of all the possible (and probable) range of faults and
> > > failures, exactly what types of "disasters" does clustering such as RAC or
> > > SFOR protect against?
> > >
> > >
> > > ken_nguyen_at_comcast.net wrote:
> > >
> > > >It has to be a very selective disaster for clustering (i.e. RAC,
> > > HACMP, etc) to provide much protection.
> > >
> > > Tim-
> > >
> > > Sorry, I don't understand what that means.
> > >
> > > Ken
> > >
> > >
> > > -------------- Original message --------------
> > > From: Tim Gorman <tim_at_evdbt.com> <tim_at_evdbt.com>
> > > It has to be a very selective disaster for clustering (i.e. RAC,
> > > HACMP, etc) to provide much protection.
> > >
> > >
> > > ken_nguyen_at_comcast.net wrote:
> > >
> > > Dan, thanks for the feedback.
> > >
> > > We're trying to protect more than just the Oracle DB. While CRS and
> > > Dataguard work well to provide HA, it doesn't take into account the Siebel,
> > > IIS, etc installs that form the entire application stack.
> > >
> > > With this solution, we're hoping to lower the TCO in the event of a
> > > disaster.
> > >
> > > Ken
> > >
> > >
> > > -------------- Original message --------------
> > > From: Dan Norris <dannorris_at_dannorris.com> <dannorris_at_dannorris.com>
> > > I have used the SFOR previously, but not on the current versions and
> > > not with 10g DB. I had no problems with the SFOR software.
> > >
> > > If I were implementing a cluster today with 10g, I wouldn't use any
> > > non-Oracle clusterware. Instead, I'd just use Oracle Clusterware as it
> > > provides all the HA you'll need for the DB. Maybe you have other reasons for
> > > using SFOR...I hope you do because I couldn't justify the investment given
> > > the current architecture.
> > >
> > > Others have posted similarly on this list and in OTN forums as well.
> > >
> > > Dan
> > >
> > > ----- Original Message ----
> > > From: Ken Nguyen <ken_nguyen_at_comcast.net> <ken_nguyen_at_comcast.net>
> > > To: oracle-l_at_freelists.org
> > > Sent: Monday, October 29, 2007 4:46:05 AM
> > > Subject: Symantec Storage Foundation for Oracle RAC
> > >
> > > We˘re looking to implement Symantec Storage Foundation HA for Oracle
> > > RAC to offer HA for our Oracle 10g RAC on RHEL.
> > >
> > >
> > >
> > > Oracle has fully certified most of the components within this Symantec
> > > solution, except for the automatic failover piece (GCO).
> > >
> > > This component is certified on all platforms except for Red Hat, could
> > > have something to do with Oracle˘s OEL initiative.
> > >
> > >
> > >
> > > Is anyone using this or any other Symantec SF products without any
> > > issues?
> > >
> > >
> > >
> > > Ken
> > >
> > > -- http://www.freelists.org/webpage/oracle-l
> >
> >
> > -- http://www.freelists.org/webpage/oracle-l
> >
> >
>
>
> --
> Andrew W. Kerber
>
> 'If at first you dont succeed, dont take up skydiving.'

--
http://www.freelists.org/webpage/oracle-l