Re: Different OS versions between Test and Prod

From: Bob Jones <email_at_me.not>
Date: Wed, 05 Apr 2006 01:53:46 GMT
Message-ID: <KqFYf.66590$dW3.6888@newssvr21.news.prodigy.com>

"DA Morgan" <damorgan_at_psoug.org> wrote in message news:1144177539.908579_at_yasure.drizzle.com...
> Bob Jones wrote:
>
>>> Is it your belief that there are no credible issues related to running
>>> one version of an operating system in test and another in production?
>>
>> As I asked earlier, give me just one example of any issue other than bugs
>> that would impact Oracle applications.
>
> And I responded with an example. A 10gR2 RAC cluster running on RedHat 3
> in test with the Grid OMA will not work with RedHat 4.
>

You've got to be kidding me. What does this have anything to do with OP's applications?

> But why say: "other than bugs?" Isn't that 99% of the reason not to do so?
>

If you read my message earlier, I am talking about Oracle bugs affecting only one of the versions of Solaris, not just any bugs. I mentioned it as an unlikely exception.

>>> That you would sign off on a migration plan where production code was
>>> never tested against the production operating system?
>>>
>>
>> What does this have to do with running the same version of OS or not? It
>> needs to be tested regardless.
>
> Exactly. It needs to be tested. So how can you have different versions
> in test and prod?

The whole point is how can you not. I ask again, give me just one example.

> Signing off on "my app works in Solaris 2.8 so
> therefore it will work in Solaris 2.10 is insane." It might work
> beautifully ... it might also crash and burn.
> --

That is also true for the same version on both systems. So what is your point. Received on Tue Apr 04 2006 - 20:53:46 CDT