<grits teeth>
Ah but MY pet peeve is being unable to find exactly what it is that the
poster is replying to if the post is at the top. My preference is for either
intersperesed comments or at the end (so I have to read the prior discussion
first as opposed to getting the reply before the point at issue)
</grits teeth>
--
Niall Litchfield
Oracle DBA
Audit Commission UK
*****************************************
Please include version and platform
and SQL where applicable
It makes life easier and increases the
likelihood of a good answer
******************************************
"Mike Ault" <mikerault_at_earthlink.net> wrote in message
news:37fab3ab.0209290912.10276fca_at_posting.google.com...
> Fellow Posters,
>
> Perhaps it is just me, but I find it really a pain to have to scroll
> to the very bottom of a tome to see your response. If I want to
> re-read the whole damn thing, I can go back to it. Place your
> responses at the top, then we can read them quickly and easily. Just
> my pet peave, and I treat it well...
>
> Mike
> "Richard Foote" <richard.foote_at_bigpond.com> wrote in message
news:<oIDl9.42496$g9.121640_at_newsfeeds.bigpond.com>...
> > Comments are all in there somewhere ...
> >
> > "Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> wrote in message
> > news:an1pc3$rpq$1$8300dec7_at_news.demon.co.uk...
> > >
> > > Have any of you received the Sept/Oct issue of Oracle Magazine ?
> >
> > No I haven't. I'm still however recovering from the July/August edition.
The
> > article on List Partitioning was hilarious (little self test: spot how
much
> > List Partitioning was actually discussed). The first inside OCP article
was,
> > how can I put it, somewhat simplistic (what's wrong with DELETE FROM
> > employee, ahhh right, no WHERE clause). Hmmm, not exactly mind
stretching
> > material and not exactly setting much of a bar for these OCPs.
> >
> > Having said that though, there was still some good stuff in the mag.
> >
> > > Have you read the article by the Certification Director for OCP ?
> > > Would you like to see a few questions, answers, and comments
> > > about DBA fundamentals.
> > >
> > > Question 1:
> > > ----------------
> > > What is the minimum number of granules allocated at SGA start-up ?
> >
> > Correct answer - who cares ? Really, from an OCP point of this is a
purely
> > theory, no practical benefit question.
> >
> > >
> > > Answer 1:
> > > ---------------
> > > The claim is three. The fact that my system
> > > always allocates at least 16MB to the shared
> > > pool alone, let alone the granules needed for the
> > > buffer and redo log when I start-up is a pity, but
> > > the answer might be right. However we then read:
> > >
> > > "On my database, I can check this by querying
> > > v$buffer_pool for LO_SETID which is 3"
> > > Why does the working data set of the default
> > > buffer pool prove anything about the number of
> > > granules ? If I have a KEEP and RECYCLE pool,
> > > the LO_SETID of the db_keep_cache_size is 1,
> > > the LO_SETID of the db_recycle_cache_size is 2.
> > >
> > > We also hear: "if the SGA_MAX_SIZE is less
> > > than 128M then the granule size is 4M, otherwise
> > > it is 16M" - perhaps he hasn't found that note on
> > > Metalink about granules of 8M on Windows 2K.
> > >
> > > Anyway - what's the point of asking someone about
> > > the notional minimum SGA of 3 granules (12MB) ?
> > > The executable is about 60M, and I'm not trying to
> > > run the thing on my cell-phone !
> > >
> > >
> > > Question 2:
> > > -----------------
> > > What four parameters most affect the SGA size ?
> > > a) sga_max_size
> > > b) shared_pool_size
> > > c) db_cache_size
> > > d) large_pool_size
> > > e) log_buffer
> >
> > Actually this question is really confusing. From one point of view, the
> > sga_max_size parameter is the one that *most* affects the SGA size.
Trying
> > to set shared_pool_size to 100M, db_cache_size to 100M, large_pool_size
to
> > 100M and log_buffer to 1 M. For some reason Oracle is not letting me. I
> > wonder if the sga_max_size being set to 64M has something to do with it
;)
> >
> > Also the non inclusion of the java_pool_size is interesting as it
> > potentially affects the SGA more so than the log_buffer.
> >
> > Again the question is somewhat academic. More beneficial questions would
be
> > how do you size and tune these parameters appropriately.
> >
> > >
> > > Well, since the SGA_MAX_SIZE is the thing that
> > > is supposed to be the absolute limit on the number
> > > of granules you can allocate to the others it clearly
> > > has to be the other four. After all, you're not really
> > > 'affecting' the size of something if all you doing is
> > > stopping it from getting any bigger !
> > >
> > > Answer 2
> > > --------------
> > > Yup, sga_max_size is the one wrong answer.
> > >
> > > We also read "The sga_max_size is determined by
> > > the size of the other memory structures". Well,
> > > yes, if you let it default - but the sga_max_size is
> > > the thing you are supposed to set (if you are being
> > > a good DBA) so that you can vary the others at need.
> > >
> > >
> > > Question 3:
> > > -----------------
> > > Which command would you execute to decrease
> > > the size of the shared pool from 50MB to 20MB ?
> > > a)....
> > > b)...
> > > c) alter system set shared_pool_size 20m;
> > > d) alter system set shared_pool_size = 20m;
> > >
> > >
> > > Answer 3:
> > > --------------
> > > (d) - perfectly correct, perfectly pointless question.
> > >
> > > alter system set shared_pool_size 20m
> > > ORA-00927: missing equal sign *
> >
> >
> > I'm sorry Jonathon but this error message has me stumped. What can it
> > possibly mean ? I've tried alter system set shared_pool_size equal 20m
and
> > the bugger still complains !! I'm beginning to think it has something to
do
> > with competitive swimming. There, pool sizes must be 50M (as currently
set)
> > or 25M for short courses. Perhaps the error message is suggesting I can
only
> > reduce the pool size to 25M ...
> >
> > >
> > > I don't care if the DBA doesn't remember that there is
> > > supposed to be an equal sign - if he can't spot the problem
> > > when Oracle gives him the error message he probably never
> > > got past filling in the application form for the interview.
> > >
> > >
> > > Question 4:
> > > -----------------
> > > Oracle Managed Files are established by setting
> > > what two of the following parameters.
> > > a) db_file_create_dest
> > > b) db_file_name_convert
> > > c) db_files
> > > d) db_create_online_log_dest_n
> > >
> > > Answer 4:
> > > ----------------
> > > (a) and (d)
> > >
> > >
> > > wait a moment - there's no such parameter as
> > > db_create_online_log_dest_n. There is
> > > db_create_online_log_dest_1,
> > > db_create_online_log_dest_2,
> > > db_create_online_log_dest_3,
> > > db_create_online_log_dest_4,
> > > db_create_online_log_dest_5,
> > > and the manual suggests that you set at least
> > > two of them. And of course the manual also
> > > tells you that you don't need to set any of them,
> > > as db_file_create_dest will do on its own.
> > >
> > >
> > > Is this hair-splitting - well if 3C is definitely wrong
> > > because of an '=' is it safe to assume that 4D is right
> > > when it has a letter instead of a number ?
> >
> > Picky picky but point taken.
> >
> >
> > >
> > > Of course, the article does continue to point out
> > > that in fact "you can set either or both". Of course
> > > the Database reference manual happens to say that
> > > there is no default value for db_create_file_dest,
> > > so I thought I'd have a go creating a database with just
> > > db_create_online_log_dest_1 - did it work ? Yes -
> > > but it built the data files in $ORACLE_HOME/dbs,
> > > which isn't really the ideal place for them.
> > >
> > >
> > > So I have learnt something today.
> >
> > I think these articles are doing some good by highlighting to anyone who
> > doesn't already know how absolutely dreadful the current OCP program
really
> > is. So Oracle deserves some credit for trying the discredit it's own OCP
> > program in it's own mag.
> >
> > The "practical" certification exam is definitely a step in the right
> > direction but as I've stated before, some significant issues still need
to
> > be addressed before anyone even pretends to take OCP seriously.
> >
> > Cheers
> >
> > Richard
> > >
> > >
> > > --
> > > Regards
> > >
> > > Jonathan Lewis
> > > http://www.jlcomp.demon.co.uk
> > >
> > > Next Seminar dates:
> > > (see http://www.jlcomp.demon.co.uk/seminar.html )
> > >
> > > ____England______September 24/26, November 12/14
> > >
> > > ____USA__________November 7/9 (MI), 19/21 (TX)
> > >
> > > The Co-operative Oracle Users' FAQ
> > > http://www.jlcomp.demon.co.uk/faq/ind_faq.html
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
Received on Wed Oct 02 2002 - 03:57:03 CDT
