Re: Oracle and High Availability solutions

I have indeed set up a cluster with 2 DEC (sorry Compaq) Alphas using OPS before, but ran into
performance problems, or at least the possibility of them !

On advice from an Oracle consultant, we were put off sharing the load between the servers
as nearly all the Oracle daemons accesses a set of common tables. The consultant
claimed that if one instance was accessing the table and the other instance attempted to access
it, that any changes (blocks) had to be written to disk before the read could start.

Is this correct ?

If so we would have had to spend a load of time 'partitioning' the database to make the best use of OPS.

In the end, the second server was used for some real-time processes, Robot and Conveyor
control, accessing the database on the main server via SQL*Net when required.

If either server failed, all the Services and IP addresses would 'bounce' to the remainin server.

This worked on a number of occassions, as we had a faulty (intermittent) SCSI bus !

BTW My original question about having only one instance active at any one time was
also related to licensing not just technical issues. Again, any experience, especially
in light of the new pricing scheme would be appreciated.

Steve Parker
Technical Consultant
LIS

                                                                                                                   
                    "Steve Orr"                                                                                    
                    <sorr_at_arzoo.c        To:     Multiple recipients of list ORACLE-L <ORACLE-L_at_fatcity.com>       
                    om>                  cc:                                                                       
                    Sent by:             Subject:     Re: Oracle and High Availability solutions                   
                    root_at_fatcity.                                                                                  
                    com                                                                                            
                                                                                                                   
                                                                                                                   
                    08/05/00                                                                                       
                    18:37                                                                                          
                    Please                                                                                         
                    respond to                                                                                     
                    ORACLE-L                                                                                       
                                                                                                                   
                                                                                                                   

I'm curious... why use cluster management software without OPS? As I understand it, with a cluster management only solution, you have to failover
to another machine by shutting down the first machine, shutting down and remounting the file system on the failover machine, and starting up Oracle on the failover machine in crash recovery mode. With OPS you just keep going
without the need for a failover. Why dedicate a second machine for failover when you can just use it in parallel? The reason I ask is because our sysadmin folks are trying to impose Veritas Cluster Manager on DBA team members who are keen on OPS.

Does anyone have any experience with Veritas Cluster Manager or know of anyone who has? Any opinions on Veritas?

TIA!!!
Steve Orr

• Original Message ----- To: Multiple recipients of list ORACLE-L <ORACLE-L_at_fatcity.com> Sent: Monday, May 08, 2000 3:44 AM

> Dear All,
>
> Consider High Availability (HA) Oracle systems, i.e UNIX and
> MC/ServiceGuard / TruCluster / HACMP etc
> where the Oracle database and binaries are held on a 'highly resilient'
> shared disk array which can be
> accessed from 2 or more servers.
>
> When either the HA Service providing Oracle is gracefully 'bounced' from
> one
> server to the other, or is restarted on an alternate server due to the
main
> server
> crashing etc, is it true to state that only one server at any one time is
> accessing
> ( or can access ) the Oracle database.
>
> Oracle parallel server is NOT being used !
>
> In all cases, 'simple' scripts are used by the HA Service to stop and
start
> Oracle,
> which normally mimic what an operator would do anyway.
>
> I would have thought that 2 sets of binaries accessing the same database
> would have disatrous consequences.
>
>
> Any thought / experience / comments please .....
>
>
>
> Thanks
>
>
> Steve Parker
> Technical Consultant
> LIS
>
> --
> Author:
> INET: Steve.Parker_at_lis.co.uk
>
> Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051
> San Diego, California -- Public Internet access / Mailing Lists
> --------------------------------------------------------------------
> To REMOVE yourself from this mailing list, send an E-Mail message
> to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
> the message BODY, include a line containing: UNSUB ORACLE-L
> (or the name of mailing list you want to be removed from). You may
> also send the HELP command for other information (like subscribing).

--
Author: Steve Orr
  INET: sorr_at_arzoo.com

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L