Items 1,2,3,5 already in place.
4 and 6 are coming as part of this.
"Daniel Morgan" <damorgan_at_x.washington.edu> wrote in message
news:1075139302.605746_at_yasure...
> Telemachus wrote:
>
> > RANT MODE ON
> > Last couple of weeks I've been trying to put together an architecture
for a
> > new highly available production internet web services and database
> >
> > W2K is the platform so 10G AS not available.
> >
> > Great I thought I can use Oracle Maximum Availability papers
> >
> >
> > I did - oh boy are there holes. Check MAA for details on IAS infrastruc
HA -
> > you won't find many.
> >
> > The database - yes - completely solid - these are technologies -
RAC,Data
> > Guard and Standby etc etc - great no problem I can live with them
> >
> > Switchover WANs network side firewalls - same here.
> >
> >
> > Even OC4J islands - yes these I can live with.
> >
> > But have any of you really tried to make the 9iAS R2 infrastructure
(SSO/
> > Metadata and so forth ) - highly available ?
> >
> > The only 9 supported solution is via clustering ! The database can't be
on
> > RAC and it must be a 9.0 database ... Why oh Why ?
> >
> >
> > (oh and unless you've patched you CAN'T CHANGE THE IP ADDRESS or the
> > HOSTNAME ! OF THE SERVER ) and even if you do patch all you can change
is
> > the IP address
> > [ this was closed as a documentation bug !!!! ]
> >
> > [ this one caught me when setting up a 9iAS box on my net for dispatch
to a
> > client site - the machine went down and joined their net and didn't
work ]
> >
> >
> >
> > Then I read the 10g docs hoping for an early reprieve - but again we
can
> > see a thread here.
> >
> > in a 10G AS DR site the machines must all have the same amount of
hardware,
> > the same config and , and the same hostnames . There's a common
something
> > running through all this that leads me to believe there were some
strange
> > decisions made in the design process for the Infra. And it's not
explained
> > properly anywhere why this must be so. (the locked-in reliance on
particular
> > hosts - which can't be changed )
> >
> > COLD FAILOVER CLUSTERS.... groan.
> >
> > Come on isn't it just an LDAP directory + a database ? .. But hang on
the
> > LDAP directory's in the db too. Do tell Oracle.
> >
> > RANT MODE OFF
> >
> > PS
> > Hi everyone nice to talk to you all again
> >
> > Telemachus.
>
> If you are truly trying to create HA then lets review what you need.
>
> 1. Two air conditioners
> 2. Power backup systems with separate diesel generators
> 3. Double routers, switches, and hubs
> 4. Double disk arrays
> 5. Double cabling
> 6. Transparent Fail-over capability for all components
>
> I doubt the App Server software is the only issue. Is someone really
> putting up the several million dollars for the supporting infrastructure?
>
> --
> Daniel Morgan
> http://www.outreach.washington.edu/ext/certificates/oad/oad_crs.asp
> http://www.outreach.washington.edu/ext/certificates/aoa/aoa_crs.asp
> damorgan_at_x.washington.edu
> (replace 'x' with a 'u' to reply)
>
Received on Tue Jan 27 2004 - 03:23:27 CST
