Re: shutdown immediate hangs

"Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message news:039j9.37528$g9.106981_at_newsfeeds.bigpond.com...
>
> "Makbo" <makbo2_at_netscape.net> wrote in message
> news:3D8D1813.4060608_at_netscape.net...
> > One "interesting thing" that can happen is if a tablespace is in hot
> > backup state, it will falsely report "media recovery required" when you
> > try to start back up. Oracle's documentation states how to clear this;
> > nevertheless it's not a fun thing to have to do in the middle of the
> > night.
>
> Anyone who issues a shutdown abort when they know their database is in the
> middle of a hot backup wants to seriously look at their administration
> skills. But that aside, as you say, 'recovery' from such a cock-up is not
> exactly difficult.
>
> >Also, under the right circumstances, I'd bet a distributed
> > transaction could experience hiccups from this.
> >
>
> Well, the instance *without* distributed transactions experiences
something
> of a 'hiccup'. But once again, if you're suggesting that shutdown abort in
> and of itself would cause some transactionally inconsistent results,
> whatever the environment, that's just not true.
>
> > Recoverable, yes, supported, yes, desireable, no.
> >
> > BTW, I've worked in environments both with daily "shutdown aborts" and
> > without, my druthers are "shutdown immediate", which under Oracle 8i I
> > have not seen hang (unlike 7.3.4). Why induce instance recovery if you
> > don't have to?
>
> I'm not saying 'go for it every time'. That's just as daft as beng scared
of
> having to use it when you need to. What this 'discussion' has been about
is
> whether it is preferable to do a shutdown abort as compared with knocking
up
> some home-produced script that does numerous 'alter system kill session'
> commands, or as compared with numerous 'kill -9' commands at the O/S
level.
>
> So, I entirely agree: it's not particularly "desirable", and an immediate
> would always be my preferred mode of shutdown (assuming I had a deadline
and
> couldn't wait for a normal to do its stuff).
>
> But that's not quite what we ended up discussing, which is this myth that
> aborts are somehow 'dangerous'.
>
> >
> > On the other hand, I'm practical -- if "shutdown immediate" hangs in my
> > environment and I can't come up with anything better in a reasonable
> > time, then "shutdown abort" it is.
> >
>
> Exactly right. A pragmatic approach, utilising all the tools Oracle makes
> available to you, is infinitely better than 'believing' something to be
> flakey when it isn't.
>
> I'm with you: when its the right thing to do, shutdown aborts are
perfectly
> reasonable things to undertake.
>
> Regards
> HJR
>
>
> > --Mark B.
> >
> > Howard J. Rogers wrote:
> >
> > > "Karen Abgarian" <abvk_at_ureach.com> wrote in message
> > > news:3D8C1067.BB75A82B_at_ureach.com...
> > >
> > [...]
> > >>There might be other interesting things related to aborting.
> > >>
> > >
> > [...]
> > >
> > > Unless you can come up with something better than "there might be
other
> > > interesting things", your "argument" holds no water. As it is, it's
> little
> > > better than voodoo and being scared at shadows.
> > >
> > [...]
> >
>

Exactly. How could it be worse than kill -9? Of course, as Howard knows, but is far too polite to point out, Oracle can recover quite easily from either. It's a great deal more robust than the doubters think.

Paul Received on Sun Sep 22 2002 - 14:33:46 CDT