Re: command line vs grid control

Noons wrote:
On Jul 11, 12:50 am, DA Morgan <damor...@xxxxxxxxx> wrote:
Most of the large firms here are using the Grid and are happy.
Among them I might mention a very large aerospace firm.

I don't have a doubt large firms are using it.
Unfortunately, large firms are NOT the majority
of db users out there. To claim that "most" are using
it because "most large firms" are, is a leap of faith
I'm not prepared to take.

Have you looked at Oracle's customer base recently? eBusiness Suite,
PeopleSoft, JD Edwards, Siebel, Retek? These organizations don't
all have classically trained DBAs in depth. For someone coming to
Oracle from SQL Server, Sybase, or Informix OEM gives them instant
access to what otherwise would take years to learn.

You know, that old story about the "expensive dba
writing scripts" is such old hat... Whoever invented that one
in Oracle marketing hasn't got a CLUE what the real work
of a dba is!
If you believe that is wholly myth I have some wonderful beachfront
property I want to sell you. It may be a myth where you are ... but
it isn't here.

Daniel, modern dbas spend the vast majority of their time
working around problems caused by appaling application
code, unstable hardware/software combos, configuring new
or replacement servers, running backups, refreshing
development and test databases, resolving interface problems
between varous server architectures and data workflow
requirements and resolving the odd performance problem
here and there. And with various databases, not just Oracle.
They also spend a LOT of time trying to work around glaring
bugs, Oracle's included.

Agreed. So you think to that burden we should add building and
testing shell scripts to perform routine tasks?

The last thing a dba wants to do is write scripts for
Oracle: they got enough in their hands already!

Amen.

Grid helps them in one of those bits above. One only.

But it could help with much more if they let it. You are totally
discounting the value of ASH, AWR, ADDM, and the advisors.

In an organization such as many where I consult there are databases
in the hundreds. Each of which requires manual patching. For some DBAs
just keeping up with this is a full-time job.

Depends on what you expect it to do. If the point is to determine
which resources are up/down then the hours is roughly 0.5 per server.
If you are talking about writing your own plugin, of course, that is
quite another matter.

Precisely.

But "undivided attention?" You must be doing something with the Grid
I'm not familiar with. Out here we have it contact us ... we don't
sit at the monitor eyes glued to the screen.

Well, for that they don't need to cough up the moolah for grid. What it does in that chapter they already do very simply and cheaply and so does any dba that has been on the job for a while.

That is not what I hear from corporate IT management. They have a very
different impression.

Where grid can help with new stuff requires the dba to sit around and
capture and examine patterns, a process that will always consume time no matter how much automation you throw at it.

Have you actually worked with OEM Grid? I don't mean the casual look on
a test box but actually worked with it for an extended period of time
managing multiple servers as part of a team with multiple DBAs?

Turn the question around. Assuming the life of the product is 5 years
how much grid control can you purchase for $1,000,000? And it doesn't
take a vacation or leave for dinner with the family or catch an attitude
in the middle of a meeting?

Of course not, but can it tango?
Sorry, couldn't resist. :-)

No but it can tell the occasional off-colour joke or two.

Enough budget is irrelevant. For-profit business exist to make a profit.
And upper management, unless you've somehow missed the offshoring trend
issue budget-irrelevant mandates with respect to FTEs.

Dan, IT budget is nowadays a minimal fraction of
the worries of upper management of large companies.
The one I work for has an IT budget that can only be
described as small change compared to what they make
elsewhere. And don't think by this I mean we spend
peanuts in IT: nothing could be less true. The post-2k
era of reducing costs is long gone, IT is not a big ticket
anymore. To try and beat that horse to death will only
ensure no one is listening...

And yet they continue to save pennies, at least in theory, by
off-shoring and, in the US, trying to circumvent caps on H1B visas
so they can bring in more people at a lower cost.

Lighten up noon ... OEM grid has its place. Perhaps not in every firm.
But in enough that it makes a big difference.

Sure, Daniel. I know what you mean. But
let me just propose this:

Oracle is going to have to do a bit more than just
whoo-in the big brass while beating up on the
dba as the "evil of IT".

Agreed.

They are up against tremendous competition in the
small to medium size shop and are losing at the rate
of knots.

I don't have the actual numbers so I can't say that this
is or is not true. What I can tell you is that I don't see
anyone looking worried about it.

Meanwhile, companies like Microslop with
their much more friendly products and pricing structures
are winning big time there.

Perhaps where you are. Where I am most companies of an size
have Oracle on-site for their line-of-business applications.
Including, though I expect the whining to begin immediately,
Microsoft, which I know for a fact, just last week, was bringing
in more people with Oracle skill sets.

You don't hear Microslop blaming dbas: they blame
products that do *not* help the dba. There is a world of
difference right there. They never tried to eliminate the
dba: they simply say the dba has life made easy with
their products.

Most SQL Server shops don't really have DBAs in the sense that
we use the word in Oracle. Want to configure blocks size? You
get your choice of 8K or 8K. Want to run a 10046 level 12?

For the vast majority of what I see here SQL Server isn't
really on the radar screen ... not just because it is SQL
Server but because of Windows. And Vista has set Microsoft
back quite a distance: It is, at least in its current iteration,
a horror story.

If you refer back to my list of what a dba does nowadays,
you see a list of things that grid and such has no hope in
heck of addressing anytime soon. But you also see
a large number of things that are incomparably easier to
achieve with the Microslop toolset. And cheaper too.

Cheaper? I'm not sure that is correct. Are you comparing Oracle
EE with Microsoft EE? I hope not. Lets compare Microsoft EE
with SE1. No amount of money in Microsoft technology will get
you RAC or TAF or Data Guard or numerous other technologies.

However I would like to drag you back to what started this thread
which was shell scripts. Written any shell scripts in Windows to
manage SQL Server lately? No anyone that has?

Any wonder Microslop's db market share has sustained
higher growth than db2 and oracle combined, for 6 years
now? And let me also state this: Oracle isn't making any
friends with their constant "the dba is evil" nonsense...

At the rate my friends in Redmond are going they may never produce
another version of a major product again. The place is growing
more and more dysfunctional with each passing year.

Chief, go outside the States and it's a Microslop world.
Oracle folks better wake up...

Not in Asia. <g> And given what I do know of Oracle's position
in the EU market I suspect they may have more to fear from
Microsoft's failures than successes.
--
Daniel A. Morgan
University of Washington
damorgan@xxxxxxxxxxxxxxxx (replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org
.

Relevant Pages