>thanks for the information. The Shareplex / Replication is planned
>from the non-RAC instance on server A to the non-RAC instance on
>Basically, we have 2 similar applications that perform similar
>functions, but we don't currently have the time or resources to
>combine them into one. One app uses OPS/RAC and the other doesn't, but
>we feel they are similar enough and small enough that we can co-exist
>them on the same server, barring any technical limitations. The
>RAC-based application has built-in failover, the non-RAC doesn't,
>hence the need for Shareplex / Replication.
Not a problem. Only concern I have is if you mean the RAC-based application has
built in failover because of RAC. That's only covering part of the failover -
machine failover to be specific. That doesn't protect you from site failure,
say for example a flood in the computer room. You'd still need to use
replication or Data Guard to have complete protection.
>> >I have 2 IBM P680 servers, clustered. Is it possible to have 1 RAC/OPS
>> >and 1 non-RAC/non-OPS instance on each server, running at the same
>> >assuming I have sufficient CPU & memory resources, and separate the
>> >correctly? Are there any limitations in 8i / 9i / OPS / RAC that would
>> >I would like to have Server A (instance A) and Server B (instance A)
>> >in a RAC
>> >or OPS configuration, and Server A (instance B) and Server B (instance
>> >B) in a
>> >non-RAC / non-OPS configuration, but running SharePlex / Replication.
>> >Is this possible, and if so, is it supported by Oracle?
>> >Roy Speaker
>>Well you have a couple of different combinations here. Let me outline each one
>> that is supported:
>> 1. both OPS and non-OPS
>> 2. both RAC and non-RAC
>> 3. OPS and non-RAC
>> 4. non-OPS and RAC
>> Proviso for all of these is that OPS/RAC is supported on the cluster you're
>> talking about. Only concern would be the resources of the machines are
>>Now add replication into the picture. Replication from Oracle is supported in
>>all 4 cases. I'm assuming here you're going to replication to another machine
>>that is not in the cluster - don't laugh at me making this explicit. I had one
>>client that has set up replication from one node of the cluster to another node
>> becasue they hadn't heard of OPS (this was in the 7.3 days). You may want to
>> ensure the replication is done from the correct node to minimize pinging.
>> Two final points. I believe (though I'm not 100% sure) that Oracle does NOT
>>support Shareplex because it uses log mining without using Oracle's LogMiner.
>>Second one - if you have the choice, don't even bother with OPS. RAC is a lot
>>easier to get scalability out of. It's possible to make OPS scale, but in some
>> cases it can be quite difficult.
>> HTH. Additions and corrections welcome.
>> SELECT standard_disclaimer, witty_remark FROM company_requirements;
HTH. Additions and corrections welcome.
SELECT standard_disclaimer, witty_remark FROM company_requirements;