This is an 8.1.7.0.0 data on Solaris 8.
I set monitoring to YES in all 14000+ tables.In the sys.dba_tab_modifications table there are 207 tables. Question: Does this mean that there are 207 tables with STALE staistics? From what I have read, that answer is suppose to be - yes.
When I run the foll
No, only the original SQL*Loader DIRECT=TRUE does that (adding blocks above
HWM), which was introduced with v7.0.x...
Since then, all direct-path (a.k.a. append) operations (including SQL*Loader
DIRECT=TRUE PARALLEL=TRUE, parallel CREATE INDEX, parallel CREATE TABLE AS
SELECT, and INSERT /*+ APPEN
Babette,
We are rolling out a *major* app with 14-16 production databases. All of the *core*
processing is in COBOL. The app has a Web front end.
COBOL won't die.
I picked up a bumper sticker, from Fujitsu's booth at Oracle World last week, that
asks "got COBOL?".
- Kirti
PS : I did COBO
Joe,
It may be unsupported, but it does work. One must be careful in saving the original
catexp.sql, and running it afterwards (after the 'tweaked' export is done). I have
been using this 'tweak' for quite some time.
Excluding user(s) from export is a fairly easy tweak.
Oracle already excludes
FWIW... Here is an interesting article by Jonathan Gennick:
http://www.onlamp.com/lpt/a/2640
- Kirti
-Original Message-
Sent: Wednesday, November 20, 2002 4:44 PM
To: Multiple recipients of list ORACLE-L
I am thinking that it must have been, since the previous email correctly
shows th
EMC Celera would be a NAS, not a SAN, no?
- Original Message -
To: "Multiple recipients of list ORACLE-L" <[EMAIL PROTECTED]>
Sent: Wednesday, November 20, 2002 7:49 AM
> We just bought EMC Cellero.
> I do not know the reasons, just that they won the contract.
>
> Yechiel Adar
> Mehish
Title: Dynamic views
We have a different business case but you might
want to look into materialized views.
- Original Message -
From:
david
hill
To: Multiple recipients of list ORACLE-L
Sent: Wednesday, November 20, 2002 1:48
PM
Subject: Dynamic views
in the past there was a way described here on the list, TOTALLY
UNSUPPORTED by tweaking one of the underlying views.
joe
Chris Stephens wrote:
i'm 99% sure that can't be done. you'll have to explicitly name all
the other users with owner=user_a,user_b,... ...then you can do an
import with
That parameter does not make CBO comes up with any smarter plans - just more likely to
choose NLJ whether its better or not. HJ can outperform all other joins in a lot of
cases if you maintain statistics judiciously, including the appropriate use of
histograms.
>>> [EMAIL PROTECTED] 11/21/02 1
Title: RE: export in full mode but exclude particular user?
i'm 99% sure that can't be done. you'll have to explicitly name all the other users with owner=user_a,user_b,... ...then you can do an import with full=y ...or flip/flop by full=y on export and owner=.. on the import.
...first
On Wed, 20 Nov 2002, Babette Turner-Underwood wrote:
> As far as staff and skills, the current job market
> has an excess of persons with PL/SQL skills.
> So I don't think the decisions was made with any
> respect to how many people will be able to maintain
> this over the next 10 years. The nu
Thanks for all the comments
I did do a little more digging.
I found out that the reason they are using Cobol
is that there are Cobol programs that have the
business logic for doing benefit calculations.
They do not want to re-write those modules.
So they are doing the entire application in Cob
Dear DBA gurus,
Has anyone got such experience when export database in full mode but exclude a
particular user?
TIA,
Chuan
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Chuan Zhang
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051 http://www.
Thanks but EVERYTHING is on the mainframe (including Oracle).
As far as I know there is no such thing as PIPEs and daemon
processes on the mainframe.
I can place skeleton JCL in a proclib. The problem is getting
Oracle on the mainframe to somehome invoke the internal reader
to submit the JCL.
---
We are using the Hitachi 9200, and have experienced no troubles in the year
or so we have had it. I have no experience with the other solutions you
mentioned. I lost the RAID configuration battle, and we have 3 databases,
one production(oltp), two development/reporting spread across a 5 disk raid
EMC has now hardware striping. The smallest stripe size is one track (1MB).
Regards,
Waleed
-Original Message-
Sent: Wednesday, November 20, 2002 5:54 PM
To: Multiple recipients of list ORACLE-L
A number of papers recommend a stripe size of 1 M (even for EMC) for volumes
containing dat
A number of papers recommend a stripe size of 1 M (even for EMC) for volumes
containing data files. I also have the following email from Eyal Aronoff of
Quest dated Nov 2000. A number of the white papers are more recent.
The reasons fo
I am thinking that it must have been, since the previous email correctly
shows that it will not execute. I do not have access to the actual db and
code yet, just paper copies of the queries that I was given to look over.
Thanks to everyone for their feedback on this. Obviously this is not an
issu
Wendy - Could this SQL have been generated by a program?
Dennis Williams
DBA
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original Message-
Sent: Wednesday, November 20, 2002 4:00 PM
To: Multiple recipients of list ORACLE-L
Knowing about how a hint works with aliases is what made me wonder abou
Title: Message
I do
not think Oracle will run the sql below. Once you have an alias for a table,
only that alias could be used as an alias (not the table
name).
Example:
Oracle9i Enterprise Edition Release 9.2.0.2.0 - ProductionWith the
Partitioning, OLAP and Oracle Data Mining optionsJS
Jack - If I can answer without stepping on David's toes, the problem with
do-it-yourself partitioning is to avoid changing the SQL. The application
may not use PL/SQL, so dynamic SQL wouldn't be a benefit.
Dennis Williams
DBA
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original Message-
[mailto:[
Knowing about how a hint works with aliases is what made me wonder about
impacts to other areas. I have never seen it mentioned in any of the
manuals I have read, and I just thought maybe someone else had experience
with this. I am trying to figure out why the developer of these queries
took the
Title: RE: Query tuning with tablename alias
but if you provide hints on such statements, you better be using aliases for hints .
Aliases are used for readability ... you either use the aliases or user tablename.column but not both ... world is already confusing enough ...
Raj
_
Wendy,
I have never heard of alias' providing either a performance gain or
reduction. One thing I haven't verified though is the impact on hints. It
has been my observation that if a table has an alias then that alias must
be used in the hint. For example:
select /*+ index(user) */ *
from ctcs
We're a development shop with a fairly robust product and some very
demanding clients, one of which would like complete error-checking and
rollback capabilities built into the delta scripts we distribute with each
new release of our product. These releases can include 100 or more separate
mini-scr
Title: RE: SQL puzzle - using where (a, b, c) in select (a, b, c from...
-Original Message-
>From: Naveen Nahata [mailto:[EMAIL PROTECTED]]
>Sent: mercredi, 20. novembre 2002 03:58
>
>Why not use some suffieintly random and crap value
> lets say '~~~CRAP~~~' to replace nulls in
Title: to_date function and NLS settings on client
Helmut,
The cleint nls_date setting can be done by
ALTER SESSION SET NLS_DATE_FORMAT =
'dd.mm.';
The currect setting can be seen from
v$nls_parameters and if the session has explicitly set it, the new value
will be shown.
Howeve
David,
You're talking about Partitioned Views, which was the predecessor of partitioned
tables. This was the drawback of partitioned views - you had to insert into the
individual tables. You can make a sql*loader script figure out which table to insert
into (if you're using sql*loader). Ot
Title: Dynamic views
Have a
job (cron or dbms_job) that fires the first day of the month to recreate the
view.
Waleed
-Original Message-From: david hill
[mailto:[EMAIL PROTECTED]]Sent: Wednesday, November 20, 2002
1:48 PMTo: Multiple recipients of list ORACLE-LSubject:
Dy
David,
How about Dynamic SQL in an Instead Of trigger?
Jack C. Applewhite
Database Administrator
Austin Independent School District
Austin, Texas
512.414.9715 (wk)
512.935.5929 (pager)
[EMAIL PROTECTED]
Wendy - I think the difference between using an alias or not is negligible.
My reasoning is that this would be easy to test (good idea if you have a
moment) and there are enough picky Oracle developers that if this was not
negligible, people would have been bragging about this as their secret
metho
Title: Dynamic views
Sure, You can write
an INSTEAD OF trigger on the view that inserts into the right table.
Arup Nanda
www.proligence.com
- Original Message -
From:
david
hill
To: Multiple recipients of list ORACLE-L
Sent: Wednesday, November 20, 2002 1:48
Hi,
I am having problem with html output from oracle reports.I have designed a
report which looks for *.jpg pictures location in the database and then from
physical location generate report as html.For this output itself converts
*.jpg to *.gif so the output comes with pictures but very shaded pi
Hi Guys...
I have an application with a remote client... The client is written in
Smalltalk and connects to the database via OCI... It establishes a
connection to the database and keeps it open... If the network link between
the remote client and the server fails, the remote client does not r
Title: Dynamic views
Hi Guys
I'm trying to see if I can get around paying oracle $50,000 for partitioning
I have a huge table and I want to partition it on date so I created 12
tables
_JAN
_FEB
And so on
But I want to create a view that will be used for inserting so it always
P
Title: Message
Hello
list.
I have recently been
tasked with trying to optimize some slow performing queries (Oracle 8.1.7) for
an application that generates reports in a data warehouse type
environment. I have noticed in most of the queries that the table names
have been aliased, but not
Dennis,
Don't get frustrated with such remarks. You are one of the most active
contributor/member of this list and atleast I like your signature 40%OCP.
Yes, I agree with you that OCP is not the final word for expertise. It
definatley force you to read most oracle material to get it certified b
On page 204 of Oracle9i RMAN Backup and Recovery (available on Amazon!) the
answer is written thusly:
When using database version 8.1.7
When using RMAN version 8.1.x
Catalog Schema requirement:
Oracle 8.1.x (or later) catalog database, with a catalog schema created by
an RMAN version that is eq
MmmThats why order by piece is there.
As the name suggests( which is very rare in oracle
naming conventions..:) j/k ) the long SQL is broken
into pieces.
so be a sport and try it !
Cheers,
RS
--- dist cash <[EMAIL PROTECTED]> wrote:
> I don't think so. V$sqltext even worse. The
> v$sqltext
I'm driving down on Monday morning.
Should be OK because the traffic system round Brum is fairly quiet and easy
to navigate during rush hour isn't it?
=)
Mike
-Original Message-
Sent: 20 November 2002 16:19
To: Multiple recipients of list ORACLE-L
I will be arriving on Sunday and am o
The best time is most probably Monday evening. There is always a gala
evening on the Tuesday night that a lot of people attend (and speakers
always get free tickets for this as well so..). Monday night, there is
always the "reception" with plenty of booze to go around - so going for a
beer and some
He he. There are 5 OCP modules. I've passed 2, hence 40%. I just thought I
should warn everyone that I'm not the expert that others on the list are.
:-) And no, I don't believe the OCP is the final word on Oracle expertise by
any means. I should probably discontinue using that signature to avoid
an
In the docs I've just checked for 9.2 Oracle recommends that
log_checkpoint_timeout, log_checkpoint_interval and fast_start_io_target are
all set to 0 if you're using fast_start_mtr_target because (as the tuning
manual says) "..setting these parameters to active values interferes with
FAST_START_MT
I don't think so. V$sqltext even worse. The v$sqltext definition
SQL> desc v$sqltext
Name Null?Type
-
ADDRESSRAW(4)
HASH_VALUE
Thanks for the follow up. I set TNS_ADMIN a different way; I will check
into this method. BTW - I mentioned the server trace because it also
provides information--sometimes more detailed--about the failing client
connection.
Debi
At 06:28 PM 11/19/2002 -0800, you wrote:
Thanks. I had run "ad
Rachna - Yes, RMAN is VERY picky about Oracle versions. What I understand is
that you are trying to use an 816 catalog to back up an 817 database. I
think that you usually must keep your catalog at the level of the target
database or higher. In other words, you could use an 817 catalog db to back
u
you are evil. :)
as far as I know, mttr overrides log_checkpoint_interval if set. That's
what it says in the docs I've looked at
I haven't had a chance to play around with it as yet though.
--- "Fink, Dan" <[EMAIL PROTECTED]> wrote:
>
> By all means, play around in their test database. See
Rachna,
you forgot one more test:
816 version of Rman against an 817 target database using an 816 version of
catalog - should work ok.
other than that, your test results are as expected. you can use a lower
version of rman against a higher version of the catalog, but the opposite is
not true.
IIRC (and it's been at least 3 years), there was a restaurant/pub on
the "island" near the conference center that was nice. Large, decent
food and excellent beer
of course, if you just want the sugar/caffeine rush, you can always go
visit the Cadbury factory
--- Connor McDonald <[EMAIL PROTECTED
I will be arriving on Sunday and am open to a get together that evening.
For those who have attended previously, what are good times?
-Original Message-
Sent: Wednesday, November 20, 2002 8:50 AM
To: Multiple recipients of list ORACLE-L
IIRC (and it's been at least 3 years), there was a
Dennis Williams
DBA, 40%OCP
^^
How does one get to be a 40% OCP? Is that like being a 40% expert?
Adam
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Donahue, Adam
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051 http://www.fatci
v$sqltext
order by piece
hth
connor
--- dist cash <[EMAIL PROTECTED]> wrote: >
>
>
> I need check what is SQL statement runnig on ORACLE.
> their have two view
> v$sql and v$sqlare. The entry "sql_text
> varchar2(1000)", it is too
> short to show SQL statement. Some of the sql
> Statemen
Title: RE: How to check/show long SQL statement?
First find the ADDRESS from v$sqlarea. Then run the following:
select sql_text from v$sqltext
where address = '8758'
order by piece;
It's ugly but all there.
Jerry Whittle
ACIFICS DBA
NCI Information Systems Inc.
[EMAIL PROTECTE
Good examples are SAP and Documentum which "migrated"
from RBO to CBO by dropping the index adjustment from
100 to 10 (thus making virtually any index look
fantastic, ie simulating rule)
Unlike them, I would recommend a little more
dilligence (as Tim's document sets out)
hth
connor
--- "Sakthi
Gillian - I love it! Dueling Instructors! Since you are still in class, why
don't you show him the manual and ask for clarification. Please check the
class notes on this. My copy is at home. Thanks.
Dennis Williams
DBA, 40%OCP
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original Message-
Sent
v$sqlare..??!
Well,
You can check v$sqltext.
Cheers,
RS
--- dist cash <[EMAIL PROTECTED]> wrote:
>
>
>
> I need check what is SQL statement runnig on ORACLE.
> their have two view
> v$sql and v$sqlare. The entry "sql_text
> varchar2(1000)", it is too
> short to show SQL statement. Some of
Joe - Based on my admittedly meager experience, I think it would work fine
for that purpose. If you use something like RMAN that doesn't interfere with
production, then speed doesn't matter.
Dennis Williams
DBA, 40%OCP
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original Message-
Sent: Tuesday,
We just bought EMC Cellero.
I do not know the reasons, just that they won the contract.
Yechiel Adar
Mehish
- Original Message -
To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
Sent: Tuesday, November 19, 2002 8:49 PM
> Hi All,
>
> We are considering the following 3 SAN st
DBAs,
I experienced following while trying to test interoperatibility
between 816 and 817.
Any comments, experiences?
Target -> 817 Catalog -> 816
rman rcvcat rman/rman@rcatqual
connect target
register database;
.
RMAN-06429: RCVCAT da
I was off yesterday and came in this morning anticipating lots of feedback
on this but now I'm surprised by the lack of response from all you seasoned
gurus. :-) I guess this is a very new thing and you aren't aware of it yet.
Anyway, here's what I've garnered so far:
There was a discussion o
You guys must be looking in ALL the wrong places!! ;)
-Original Message-
Robert
Sent: 20 November 2002 14:04
To: Multiple recipients of list ORACLE-L
>> I'm in - when you consider the wealth of nightlife in Birmingham ...
LOL you must have been in a different Birmingham than I was C
We just bought EMC Cellero.
I do not know the reasons, just that they won the contract.
Yechiel Adar
Mehish
- Original Message -
To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
Sent: Tuesday, November 19, 2002 8:49 PM
> Hi All,
>
> We are considering the following 3 SAN st
Use v$sqltext.
select sql_text from v$sqltext where address = xxx order by piece;
dist cash wrote:
>
> I need check what is SQL statement runnig on ORACLE. their have two view
> v$sql and v$sqlare. The entry "sql_text varchar2(1000)", it is too
> short to show SQL statement. Some of th
I am trying to install Oracle 7.3.4 on a MS2000 machine. I need the loader
utilities to run against an old instance of Oracle Financials. I am getting
a "obackup.vrf OS_ERROR". I have done this before and I did not get the
error. Can anyone help?
R. Smith
If you are not the intended recipient
Vivek,
As John K Suggested, take a look at Tim's paper
"The Search for Intelligent Life in the Cost-Based
Optimizer". He discussed the impact of these settings.
http://www.evergreen-database.com/
Gaja also discusses these parameters and their
possibly optimal values in his paper. You can find
tha
Title: to_date function and NLS settings on client
Helmut,
your format is wrong, it should be:
to_date('01.10.1950','DD.MM.')
Igor Neyman, OCP DBA[EMAIL PROTECTED]
- Original Message -
From:
Daiminger, Helmut
To: Multiple recipients of list ORACLE-L
Sent
Title: to_date function and NLS settings on client
>How coem this does not work although I
explicity specify a format mask?
Maybe
because the mask is wrong... ;)
try:
to_date('01.10.1950','DD.MM.')
instead of:
to_date('01.10.1950','DD-MM-');
or the
other way around!
/Johan
--
http://www.google.com/
and search for: vmtune depth queue
Rich
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]] On Behalf Of VIVEK_SHARMA
Sent: Wednesday, November 20, 2002
12:40 AM
To: Multiple recipients of list
ORACLE-L
Subject: RE: AIX vs Solaris
I need check what is SQL statement runnig on ORACLE. their have two view
v$sql and v$sqlare. The entry "sql_text varchar2(1000)", it is too
short to show SQL statement. Some of the sql Statement we use > 3000
characters. Does their has way to check/show long SQL statement?
Thanks.
>> I'm in - when you consider the wealth of nightlife in Birmingham ...
LOL you must have been in a different Birmingham than I was Connor! :-))
Certainly not the nightlife of London, to be sure.
The Christmas fair (or whatever it's called) was pretty nice last year.
RF
Robert G. Freeman -
Charlie,
Cary provided some really good info on some things to look at, but I'm
going on a really wild hunch here (well, okay, so I actually have a few
different thoughts on it and not just one hunch)...so some further
information would be needed.
INVENTORY_LOCATION sounds like a good asso
What would be an Optimal Reduced Value of OPTIMIZER_INDEX_COST_ADJ
to make it favour index searches ?
Default Value is 100.
Thanks
-Original Message-
Sent: Wednesday, November 20, 2002 4:54 PM
To: Multiple recipients of list ORACLE-L
One quick hack is to set the "optimizer_..." para
It'd be good to grab something to eat and wind down after the 'rigours' of
the day.
Needn't be anything flash as I'll probably be in more dire need of a beer
than anything else.
Mike
-Original Message-
Sent: 20 November 2002 12:19
To: Multiple recipients of list ORACLE-L
I would sugge
Yes , did some small benchmarks of a Batch process running for a Few minutes & Loading
the CPUs by 50 %
Database Exists on Veritas Filesystems (VXFS)
truss of the process Both application & Database Calls made
Run with Value of vxio:vol_maxio = 512 ( = 256 K [Default] )
-
I would suggest anywhere along Broad Street (opposite the "front" of the
ICC), or anywhere in Brindly Place (across the canal at the "back" of the
ICC). There are numerous bars/cafes/restaurants around the Broad Street
area..
I suppose the question is - do we want to have a meal, or just get toget
Title: to_date function and NLS settings on client
Hello!
We have several Windows clients connecting to our db-server. Every client issues the statement:
select from thabe where date = to_date('01.10.1950','DD-MM-');
Some clients return the correct number of rows where as others
Title: RE: SQL puzzle - using where (a, b, c) in select (a, b, c from...) an
Jacques,
Why not use some suffieintly random and crap value lets say
'~~~CRAP~~~' to replace nulls in NVL or DECODE for the query to work
with Varchar columns? I think one can be reasonably sure that such a
When the max value of the 4 bytes or 8 bytes have been reached the
values may become negative, if oracle keeps on adding to them.
Anjo.
-Original Message-
Nahata
Sent: Wednesday, November 20, 2002 11:34 AM
To: Multiple recipients of list ORACLE-L
Thanx Anjo,
Can you elaborate on 'values
Connor,
I see you're presenting this year. I'll try to catch those sessions.
Assuming that you know the format of the event better than I do (not
difficult) have you any suggestions for good meeting places/times?
Mike
-Original Message-
Sent: 20 November 2002 11:19
To: Multiple recipients
One quick hack is to set the "optimizer_..." params to
make indexes look super-wonderful, which then gives
you a rule-based "look and feel" with hopefully a
little smarter selection of indexes as a bonus
hth
connor
--- VIVEK_SHARMA <[EMAIL PROTECTED]> wrote: >
> Consider using OUTLINES for Fixe
I'm in - when you consider the wealth of nightlife in
Birmingham (yeah right!) then I'm sure we can find
somewhere to go :-)
--- Mark Leith <[EMAIL PROTECTED]> wrote: > I'm
interested! It IS always nice to put names to
> faces..
>
> Any ideas for when/where?
>
> Mark
>
> -Original Message-
Tim -- that adds new blocks above the HWM?
I wonder if any command that adds new blocks (vs inserting rows into
existing ones) automatically creates the new blocks as TEMP segments.
Logically it makes sense but I wonder if it's documented anywhere.
--- Tim Gorman <[EMAIL PROTECTED]> wrote:
Thanx Anjo,
Can you elaborate on 'values may wrap'?
Regards
Naveen
-Original Message-
Sent: Wednesday, November 20, 2002 2:34 PM
To: Multiple recipients of list ORACLE-L
1) values may wrap (and there are some obscure bugs that cause negative
values)
2) You want to get the block in mo
VIVEK_SHARMA,
Hi, did you have benchmark result or other whitepaper talking about
this thoery?
For db file sequential read, it is single block read to index and then
to table, so i think enlarge the maxio size of the os won't help?(I just guess, did
not test it).
Orr, Steve,
I am using Xeon MP as well on my DELL 6650 , and i have two nodes. I
will patch my database next evening and later turn on the logical cpu feature on one
node and leave the other node as it is.
I will compare the cpu load/application responce time change and gi
I'm interested! It IS always nice to put names to faces..
Any ideas for when/where?
Mark
-Original Message-
Mike (NESL-IT)
Sent: 20 November 2002 09:13
To: Multiple recipients of list ORACLE-L
I'll be attending and it would be nice to put faces to some of the names on
this list.
Mike
Heh, naw it was an early morning pre-coffee reply. So probably nothing to
write home about =)
Mike
-Original Message-
Sorry Mike. I was going from memory. I'm sure your reply was better than
mine, not hard to do.
Dennis Williams
DBA, 40%OCP
Lifetouch, Inc.
[EMAIL PROTECTED]
--
Plea
I'll be attending and it would be nice to put faces to some of the names on
this list.
Mike
-Original Message-
Sent: 20 November 2002 07:28
To: Multiple recipients of list ORACLE-L
I will be at UKOUG and so will be another of other people of the Oaktable
network.
Anjo.
-Or
1) values may wrap (and there are some obscure bugs that cause negative
values)
2) You want to get the block in mode CR and there is actually no work needed
(like cleanout or rollback) to get to that mode.
Anjo.
On Tuesday 19 November 2002 21:34, you wrote:
> Hi All,
>
> There is one report wh
On Tuesday 19 November 2002 22:03, you wrote:
> Qs What is the Cause in particular (or in General) of Time Difference
> between "cpu" & "elapsed" Columns in the following Query ?
>
e = c + wait time (of anykind) (+ rounding errors)
> Qs Is there Any Scope for improvement in the following Query ?
>
For "db file sequential read wait" Add Following to /etc/system
On UFS
Setting maxphys=8388608
On VXFS
set vxio:vol_maxio = 16384 ( implies 8MB )
-Original Message-
[mailto:[EMAIL PROTECTED]]
Sent: Tuesday, November 19, 2002 8:24 PM
To: Multiple recipients of list ORACLE-L
I'm search
I will be at UKOUG and so will be another
of other people of the Oaktable network.
Anjo.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]] On Behalf Of Fink,
Dan
Sent: Tuesday, November 19, 2002
9:19 PM
To: Multiple recipients of list
ORACLE-L
Subject: O
It did that for the longest time. It
creates a temp segment so that if the create table table
fails near the end, there doesn’t have to be a complete rollback. SMON
can come along and cleanup the segment. If the create table as has completed,
the temp segment will be converted to a ordinar
93 matches
Mail list logo