Re: command line vs grid control

On Jul 6, 12:24 pm, "GD" <goran99_ma..._at_yahoo.com> wrote:
> <sybra..._at_hccnet.nl> wrote in message
>
> news:lh2t83dtpanm7ho945l2r0gcpg816rko40_at_4ax.com...
>
>
>
>
>
> > On Fri, 06 Jul 2007 16:37:08 +0200, GD <goran99_rem..._at_yahoo.com>
> > wrote:
>
> >>How do you invoke rman if you use it? Why wouldn't you use a shell
> >>script to do that?
>
> > two words
> > hardcoded passwords.
> > At one of those users has SYSDBA privilege by design
>
> >>> Take the tool of your choice and see whether you can:
>
> >>> Use block change tracking
> >>> Perform block media recovery
>
> >>> The answer in both cases is no.
>
> >>Probably true, but why would you terminate people who use
> >>perl/shell/whatever scripts to invoke whatever backup tool?
>
> >>Regards
>
> > Why would you continue to use a 'backup tool' which isn't integrated
> > with Oracle at all, which does generate *more* redo during backup,
> > which doesn't automate your backup and which doesn't automate your
> > restore/ recovery?
> > Rman has been around since 1998. People who still don't know about it
> > need either to be trained or shown to the door of unemployment.
>
> I almost completely agree with you, except that there are companies that use
> storage mirroring and don't need to use rman.

There are also companies that depend entirely on storage mirroring... and then post here asking how to recover their noarchivelog db when the 100000 hour mtbf turns out to be a mean mean.

> If you use storage vendor replication/mirror/clone features (which are
> supported both by Oracle and your favorite storage vendor), you probably
> have to use some kind of shell scripting (or whatever automation) to do the
> backup job.
>
> Even if you use rman, I really don't quite understand why "anyone using
> shell/perl scripts to perform a backup should be trained or terminated". Is
> it bad practice to use rman from within a shell script?

Depends. More commonly the bad practice is in design or procedures, especially underestimating the value of redundancy. I think Daniel is overstating the case when older versions and smaller sites are included, but his point about being able to bring stuff back 4+ years on is an important one.

jg

--
@home.com is bogus.
Or seven years... http://www.signonsandiego.com/uniontrib/20070706/news_1b6prgn.html