Re: Benefits of RMAN

From: Mark J. Bobak <mark_at_bobak.net>
Date: Fri, 09 Jul 2004 02:39:56 -0400
Message-ID: <pan.2004.07.09.06.39.56.628364@bobak.net>

On Thu, 08 Jul 2004 16:36:40 -0500, JEDIDIAH wrote:

> On 2004-07-07, Mark J. Bobak <mark_at_bobak.net> wrote:

>> On Thu, 01 Jul 2004 17:48:11 -0500, JEDIDIAH wrote:
>>
>> (Text snipped)
>>>
>>> What Oracle is pushing us into violates "the Unix way of doing things".
>>> Since Oracle's core market is with Unix shops, this is not something
>>> to trivialize. Disasters impose considerable burdens on an enterprise
>>> without making them also contemplate whether or not RMAN will make it
>>> through in one piece.
>>
>> Excuse me? "...whether or not RMAN will make it through in one piece."??
>> If that's truly your opinion, then I can only assume you have either:
>> 1.) Never used RMAN. (Please, try it. Set up a throw away test

> 
> 	I seem to vaguely recall comparing you with a Microsoft shill
> 	before. I think I will renew that observation if you choose
> 	to trivialize those that have experiences different than your
> 	own.
> 

"Microsoft Shill"? Wow, considering that I have no Microsoft products installed on personal systems, and only run it on my desktop at work cause I have to, that's an interesting statement. Also, I don't remember any previous exchanges with you, and I wonder if it's possible you've mixed me up with someone else? (Before yesterday, it's been quite some time since I posted anything at all in this newsgroup.) Also, it was not my intention to "trivialize" anyone's opinion, only to express mine, that being "RMAN is a good thing". I apologize if I came off sounding that way.

> 	That's not a matter of "opinion". It's a matter of experience. One
> 	wonders if you've ever been in a position to do disaster recovery ever
> 	(for play or for real).
> 

Yes, I've done disaster recoveries, both real and simulated. I've done them with conventional backups as well as RMAN backups.

>> database to practice on! It really is very good at what it does.) OR
>> 2.) You looked at it, but didn't really put the time into it. RMAN
>> has a lot of annoying new lingo and terminology. If you're used to
>> conventional backups, it cant bea bit daunting. Bite the bullet, read
>> the docs, pick up the Freeman book or the Velpuri book. It's worth the
>> journey.

> 
> 	This is a RED HERRING on the subject of RMAN's trustability.
> 

Sorry, I have to disagree there. Being comfortable with RMAN, and it's terminology will certainly affect someone's trustability of RMAN, since trustability is about a person's ability to trust. People tend to not trust things that they are not comfortable with. I WOULD say that it's a red herring on the subject of RMAN's reliability, as reliability (or lack thereof) is not dependent on the comfort zone of the user. Either RMAN fundamentally does the right things and is reliable or doesn't and is not reliable.

>


>>> The cost benefit analysis doesn't necessarily work out in Oracle's

>>> favor even for large shops that might actually have thousands of files
>>> to manage.
>>
>> Now who are you kidding? That's what RMAN does *best*! It manages
>> lots and lots of files and keeps track of which versions of which files
>> go with what databases and what archive logs you'll need and where they
>> are. That's RMAN. It's what it does.

> 
> Enterprise backup managers already do this.
> 
> What will RMAN do if you happen to be missing a tape during a disaster?
> 

What will an enterprise backup manager do if you happen to be missing a tape? Also, consider that most folks implement an enterprise backup manager along with RMAN. Many vendors today supply the media management library (MML) that allows you to directly interface RMAN with the various EBMs. So, it's not a mutually exclusive proposition.

> 

>> Try it, you'll like it....;-)

>
> It's entirely too proprietary.

Well, you're certainly entitled to your opinion. I'll leave you with one last thought. With the advent of ASM in 10g, RMAN is the only game in town. Now, granted, you aren't required to use ASM, at least in 10g. What about future versions? I have no knowledge of Oracle's future plans, but the writing does seem to be on the wall. And if, in fact, that ends up being the long-term direction, you've certainly got plenty of time yet. But, I think ultimately, that's the direction we will all end up going in.

-Mark Received on Fri Jul 09 2004 - 01:39:56 CDT