Re: v$session

On Feb 5, 4:18 pm, bdbafh <bdb..._at_gmail.com> wrote:
> On Feb 5, 4:45 pm, joel garry <joel-ga..._at_home.com> wrote:
>
>
>
>
>
> > On Feb 5, 12:59 pm, Ben <benal..._at_yahoo.com> wrote:
>
> > > On Feb 5, 2:38 pm, joel garry <joel-ga..._at_home.com> wrote:
>
> > > > On Feb 5, 11:14 am, Ben <benal..._at_yahoo.com> wrote:
>
> > > > > 10.2.0.2 ent ed
>
> > > > > what underlying table is v_$session built on?
>
> > > > > create or replace public synonym v$session for v_$session;
> > > > > create or replace view v_$session as select * from v$session;
>
> > > > > These are the only ddl statements I can find related to it, they
> > > > > reference each other.
>
> > > > You can figure it out like this:
>
> > > > desc v$fixed_view_definition
>
> > > > select view_name from v$fixed_view_definition
> > > >   where view_name like '%SESS%';
>
> > > > SELECT view_definition FROM v$fixed_view_definition WHERE view_name='GV
> > > > $SESSION';
>
> > > > jg
> > > > --
> > > > @home.com is bogus.
> > > > What's in your database?http://www.mailonsunday.co.uk/pages/live/articles/news/news.html?in_a...
>
> > > Thank you very much.
> > > So have you ever updated x$ksuse or x$ksled ?
>
> > No, I don't ever mess with x$ tables unless Oracle specifically tells
> > me to.  They have never told me to - such things tend to void
> > support.  Even just looking at such things directly requires
> > understanding beyond anything supported.  Several authors have
> > published some interesting things about them.  But you're on your own,
> > and even the gurus have bought large systems down with simple
> > queries.  Educational, but not politic.
>
> > jg
> > --
> > @home.com is bogus.http://www.nydailynews.com/news/2008/01/27/2008-01-27_empire_state_bu...
>
> Joel,
>
> Okay. Here's one for you.
> I"m at the Hotsos Performance Tuning Symposium (probably 2003), the

Man, I so want to go. Not going this year to anything now that I'm a perm employee<sigh>. One of the subsidiaries isn't far from Collaborate. <big sigh>

> add-on dy with Steve Adams. Someone (probably Mark Bobak) mentions
> that Steve is  connected as sys and not with an unprivileged account.
> There was some relevance as to how something would run differently
> while connected as user sys.
>
> Steve's response was - that he always connected as sys.
> That's typically the schema that he was working in.
> And for research and development of what he was working on - on a
> single user database on a laptop, it was appropriate. Elsewhere, it
> most likely would not be.

And certainly for newbies that's way out of line on production systems. Personally I would encourage anyone to learn and dig as deep as they can with their own machines, Steve's stuff is classic, though I can see why he wouldn't want to update it nowadays.

I see that sort of research is called gray box: http://catless.ncl.ac.uk/Risks/25.03.html#subj18

>
> I faintly recall having to restore and re-run an upgrade of a dev db
> that had the ixora views installed into it. those were not something
> that you'd want to leave lying around come upgrade time. really hosed
> things.

Funny how that happens :-)

>
> Highly informative session, though. lots of "ums" and block dumps.
> Didn't keep track of which had a higher count, but there were well
> over 200 block dumps.
> Never really did fully implement the use of single table hash
> clusters. Was really fired up to do so after the session.

Definitely a risk of a good conference: enthusiastically wanting to ring every bell and blow every whistle.

>
> Maybe in 11g.
>

The wet blanket of "productivity" always mutes the bells and whistles.

jg

--
@home.com is bogus.
Bluetooth: new device found: Airbus A310  http://www.heise.de/ct/schlagseite/03/01/gross.jpg