- --- Original Message --- -
From: liujd [EMAIL PROTECTED]
To: Multiple recipients of list ORACLE-L
[EMAIL PROTECTED]
Sent: Tue, 25 Mar 2003 22:53:44
i write two function .
one is login function as login(char * password,int
contextId).this
function initialize the context
- --- Original Message --- -
From: liujd [EMAIL PROTECTED]
To: Multiple recipients of list ORACLE-L
[EMAIL PROTECTED]
Sent: Tue, 25 Mar 2003 22:53:44
i write two function .
one is login function as login(char * password,int
contextId).this
function initialize the context
Personally, I use a modification of Tom Kyte's sql*plus setup
from Expert One-On-One Oracle (book publisher bankrupt):
-- Connect.sql
CONNECT1
UNDEFINE 1
@@Login.sql
SETTERMOUTOFF
SAVEafiedt.bufREPLACE
DEFINEprompt_global_name = SQL
COLUMNprompt_globalNEW_VALUE
Thanks a lot guys.it's solved.
i found that a developer had added a column w/o my knowledge.
:(
Regards,
Jp.
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Prem Khanna J
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051
Hallo,
I have this sequence,
CREATE SEQUENCE SYS_STORE_ID
START WITH 1
INCREMENT BY 1
NOMAXVALUE
NOCYCLE
CACHE 10;
I am trying to run this script,
sys_store_id.nextval:
but get this errormessage:
unknown command beginning SYS_STORE_... - rest of line ignored.
Depending on the OS, Microsoft has different CPU counts. The Windows
2000 family
counts each hyper thread as 1 CPU. In contrast, Windows XP and Windows
2003 Server
count each CPU instead of each hyper thread. Thus, the task manager in
Windows XP
Professional show 4 CPU's (on a dual XEON CPU
--_=_NextPart_001_01C2F37A.FBC3FB80
Content-Type: text/plain; charset=iso-8859-1
We too have been down this route - our experiences may just be relevant to
you, Paula - and others.
From an early stage we recognised the crucial importance of getting the
design RIGHT. That meant at least TNF.
Thank you for your answer.
I run the program that I set error in it previousely
Then I debug it and find the error code is -2134
Now I find a weird thing.
When I connect the database and then disconnect .
After disconnection, The function getGroupInfo() can handle the error
well in the first
Hi all
I have a question which is probably easy for some of you
but it is 21:00 and my mind is not working. Would appreciate
it if someone can point me in the right direction. version is 8i
so no ANSI OUTER JOIN.
I have 4 tables:
master_tab, ID is PK
detail_tab_A, ID is FK
select sys_store_id.nextval from dual;
[EMAIL PROTECTED] wrote:
Hallo,
I have this sequence,
CREATE SEQUENCE SYS_STORE_ID
START WITH 1
INCREMENT BY 1
NOMAXVALUE
NOCYCLE
CACHE 10;
I am trying to run this script,
sys_store_id.nextval:
but get this
hi list
Is there somebody on this wonderfull list considering to move away from the
mainframe computer with
his old RDMS to another platform (MS or UNIX) with oracle/sqlserver ?
What are your scenarios : Big bang or keep 2 database in synch (how ?)
Anyone with Unisys Clearpath systems with this
select sys_store_id.nextval from dual; should solve the problem.
sys_store_id.nextval is not a sqlplus command. It can be used as
select List-item in a sql statement like the one in my example. A
'select list-item' is normally a column name from the table in the from
clause in the sql statement.
Hi,
Try something like
select id, numberrowsA, numberrowsB ..
from master_tab m,
(select id, count(*) numberrowsa
from detail_tabA a
where m.id = a.id
group by a.id) da,
(select id, count(*) numberrowsb
from detail_tabb b
where m.id = b.id
group by b.id) db,
...
where
da.id = m.id
and
Hi all,
After migration of reports to 9iDS
release 2,
we cannot directsend a Landscape
reports to printer
Printer assumes it a Portrait
report.
Did anyone face this problem
?
Regards,Salaheldin
Aboali--Senior Software Developer
Management Information
You should:
select sys_store_id.nextval from dual;
or similar to get a value from sequence.
Br.Jorma
Hallo,
I have this sequence,
CREATE SEQUENCE SYS_STORE_ID
START WITH 1
INCREMENT BY 1
NOMAXVALUE
NOCYCLE
CACHE 10;
I am trying to run this script,
FYI -
Following on the point picked up by Darrell Landrum
below, and my comment about global histograms in 8i,
I decided to test a workaround:
Could you export a histogram from a non-partitioned table,
then import it to a partitioned table ? If so, would a query
against the
Oracle corp provides some migration assistance. In my experience,
database migration is easier part ( in terms of man days, technical
difficulties , testing etc.). Application conversion/migration could be
tougher and more time consuming depending on the decision you have made
( complete rewrite/
-Original Message-From: Paulo Gomes Sent:
quarta-feira, 26 de Março de 2003 11:17To:Subject:
events number and meaning
Hi guys.
Where can i find a list of
Oracle Db (8i, 9i and 9iR2) events and menning?
and by the way how can i fire a
Stored procedure if a event (ex.: Shutdown
Good day to
everyone
How do I find under which thread is my
current session running in 9i RAC?
Vladimir
Barac
i write two function .
one is login function as login(char * password,int contextId).
This function initialize the context environment .and login
database.
two is getGroupInfo(char * groupName,int contextId);
this function is as his name to get one group information.And
refer to
Hemant - If you've got to rework your backup script anyway (and retest it),
why not consider switching to RMAN at this time? In my experience,
negligible interference with production during a hot backup.
Dennis Williams
DBA, 40%OCP, 100% DBA
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original
you can find the list of
events under $ORACLE_HOME/rdbms/mesg/oraus.msg.
babu
- Original Message -
From:
Paulo Gomes
To: Multiple recipients of list ORACLE-L
Sent: Wednesday, March 26, 2003 7:08
AM
Subject: FW: events number and
meaning
Thanks Jonathon,
I would never have thought to try this. Excellent information to
know!
(I did, however suspect the manual's statement on query predicate, etc.
would be found to not be 100% true 100% of the time.)
Thanks!
[EMAIL PROTECTED] 03/26/03 05:43AM
FYI -
Following on the point
2000 only shows 2 CPUs because it cannot take advantage hyperthreading.
From what I have read (Intel's website, some others), hyperthreading
requires both BIOS and OS support to be utilized. Currently, XP and the
newer Linux releases are all that I have seen that will take advantage of
it.
My
G.g.
If I ever heard a question that merits IT DEPENDS, this is one.
I think Pradip has provided excellent advice.
A lot depends on what is running on the mainframe. What are the critical
business processes that depend on it? I think the more the processes on the
mainframe are
On Tue, Mar 25, 2003 at 04:58:53PM -0800, Jacques Kilchoer wrote:
Every Dog Has His Day! And today's mine.
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Jacques, this indeed an exquisite use of owa_pattern!
--
Thanks everyone, for your suggestions, very
Good luck. We've had nothing but problems trying to use that config. As
far as we could tell, the DB was fine. It was the external Java stuff that
didn't work, specifically OMS. It would be fine for a connect or two, then
crash, then not want to restart.
And of course, since there are no FREE
Hallo,
select sys_store_id.nextval from dual;
Cheers
-Original Message-
Sent: Wednesday, March 26, 2003 4:19 AM
To: Multiple recipients of list ORACLE-L
Hallo,
I have this sequence,
CREATE SEQUENCE SYS_STORE_ID
START WITH 1
INCREMENT BY 1
NOMAXVALUE
NOCYCLE
Robert - I stand in awe of your insights (as always). Excellent suggestions.
Thank you.
I keep learning that the key to DBA success lies at least as much in the
people skills ad the technical skills. There is a new book out, The Art and
Science of Oracle Performance Tuning by Christopher Lawson.
thanks
regards
PG
-Original Message-From: Babu Nagarajan
[mailto:[EMAIL PROTECTED]Sent: quarta-feira, 26 de Março de
2003 12:59To: Multiple recipients of list
ORACLE-LSubject: Re: events number and meaning
you can find the list of
events under
How do i see the info
inside???
Regards
PG
-Original Message-From: Babu Nagarajan
[mailto:[EMAIL PROTECTED]Sent: quarta-feira, 26 de Março de
2003 12:59To: Multiple recipients of list
ORACLE-LSubject: Re: events number and meaning
you can find the list of
events under
We want true server-less backups. I know that RMAN's overhead is [much] lower
than that of the ALTER TABLESPACE BEGIN/END BACKUP but the
Snapshot method in a SAN [akin to splitting and resilivering a mirror, in
some ways]
should be much faster. The database is in BACKUP mode only for the time
OPS 8.0.6.3 on Sun Nodes. We have this currently set to 0, and Oracle has
recommended that we leave it at the default of 700, while making a general
recommendation to improve the performance of this database. But from what I
have gathered so far from Metalink, this would not suit us, since,
Title: RE: Oracle Licences Req on Intel Xeon CPU
Yes Redhat Advanced Server 2.1 does support hyperthreading.
Its not to bad either. I ran some tests and we get about a 15% increase in performance with hyperthreading turned on.
-Original Message-
From: Michael Brown [mailto:[EMAIL
-Is 8:50 AM too early for beer?
Not in Wisconsin!!
-Original Message-
Sent: Wednesday, March 26, 2003 8:54 AM
To: Multiple recipients of list ORACLE-L
Good luck. We've had nothing but problems trying to use that config. As
far as we could tell, the DB was fine. It was the external
It used to be that all error messages between 1 and 10999 were events
... but I see some more in 21xxx, 23xxx range too in 9i.
Raj
-
Rajendra dot Jamadagni at espn dot
com Any views expressed here are
strictly personal. QOTD:
Here is an old email from Kirti that can help you.
Hi Roy,
Planning, planning and planning is required to make this a piece-of-cake..
This one is a rather long one.. Sorry..
Prepare scripts to build tables, indexes and constraints etc.. Prebuild the
tables in the target database. I am
Listers,
We are in our first year of production and our transactions are growing
in volumes unforeseen. Our management team, our customers, and our
clients are most interested in analyzing historical trends so they can
better predict future trends. At the moment we use ad-hoc reporting at
The problem is that select statements will find rows not to be there
because the snapshot scn is back in time, but as soons a you start working
on current blocks, there is no problem. Shipping the current block will send
also some scn across potentially (if needed , look 'redo ordering marks' in
Title: alter system dump logfile ... question
I am trying to find who deleted certain rows fro a table. So I dumped the right logfile with 'layer 11 opcode 3' option. I can see the information about deleted rows, but where is it stored .. whodunit??
Here is a partial listing ... the objectid
Title: Re: Amt of Training to be considered "up to date"
A
SQL7/SQL2K Filegroup is equivalent to an Oracle tablespace.
-Original Message-From: Boivin,
Patrice J [mailto:[EMAIL PROTECTED]Sent: 25 March 2003
20:04To: Multiple recipients of list ORACLE-LSubject: OT
SQL Server 2000
Why do you have to suspend?
--
Jeremiah Wilton
http://www.speakeasy.net/~jwilton
On Wed, 26 Mar 2003, Hemant K Chitale wrote:
We want true server-less backups. I know that RMAN's overhead is [much] lower
than that of the ALTER TABLESPACE BEGIN/END BACKUP but the
Snapshot method in a SAN
1. Assuming Wait Events:
Please check the
Database Server Reference Guide for the respective release. Also check www.oraperf.com, there a list of wait events
by Oracle Versions. Free membership is required.
2. Using triggers at database
level.
for
example:
create or replace trigger
Tony - 8i does have outer joins, the syntax is (+) - intuitive right?
I haven't run this query, but it might give some ideas. Since nobody has
responded yet, maybe it'll provoke the experts ;-)
select
r1.id, r2.count(*), r3.count(*), r4.count(*)
from
(select id from master_tab) r1,
Title: alter system dump logfile ... question
I
believe you'll need to use LogMiner.
Dick
Goulet
-Original Message-From: Jamadagni, Rajendra
[mailto:[EMAIL PROTECTED]Sent: Wednesday, March 26, 2003
11:14 AMTo: Multiple recipients of list ORACLE-LSubject:
alter system dump
Anybody take this course before. Any comments, good or bad. Thanks
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Tracy Rahmlow
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San Diego, California-- Mailing
Title: alter system dump logfile ... question
Dick,
I tried various methods ... v$logmnr_contents has 0 rows ... that's why
the hard way of doing things ... as usual metablink is little low of details
about v$logmnr_contents containing no rows.
Egads ... I am finally back on the list after
Hi,
There are some DW ressources out there :
www.dw-institute.com
www.dmrewiew.com
www.olapcouncil.org
www.redbooks.ibm.com/pubs/pdfs/redbooks/sg242238.pdf (Excellent !)
Books by Ralph Kimball (www.ralphkimball.com) and Bill Inmon
For Oracle DW only, look for the book from Tim Gorman and Gary
I have not attended the clinic, but 3 DBAs from my company attended the
recent Hotsos Symposium and loved every second of it. Based on that,
I'm trying to make plans to attend one of the future clinics and
certainly intend to go to the Symposium in Vegas next year!
(Side note for the Hotsos reps
Saira
Yes, typically an OLTP system has no history. It has only a now. For
historical trends you should move data to a data warehouse. This will free
up processing resources as well.
I would start by visiting Ralph Kimball's web site
http://www.ralphkimball.com. He has many articles posted that
Title: RE: Import to Oracle9i extremely slow
We also ...
1. created tables first ... with no indexes and no constraints.
2. Prepared a constraints script
3. drop all triggers.
4. Imported data in multiple streams (in our case 6 import processes on each node X 2 nodes) with different tables
Title: global cache CR requests and _fairness_threshold parameter
In one of our test, we say lots of waits on global cache cr request ... I haven't been able to find much except that in certain cases setting _fairness_threashold can be set to 1, default is 4.
In our test db, this (setting
I would be a little surprised if the problem relates to
dynamic memory management, or the use of the
db_cache_size parameter.
However, it is possible that you managed to
change the number of db block buffers available
quite dramatically - which could have an effect
on the size at which a table
Title: veritas backup via RMAN troubles
AIX 5.1, oracle 9ir2(9.2.0.1), veritas 4.5
followed instructions for installing veritas on disk, got the agent, did NOT relink oracle executable, but ran the oracle install script so the link from /usr/openv/netbackup/bin/libobk.a64 is linked to
Tony,
this works also:
select distinct r1.id, count(r2.id),count(r3.id),count(r4.id)
from r1, r2, r3, r4
where r2.id (+) = r1.id
and r3.id (+) = r1.id
and r4.id (+) = r1.id
group by r1.id
Tom Mercadante
Oracle Certified Professional
-Original Message-
Sent: Wednesday, March 26,
I have got a set of datafiles backed up from the client but he didn't send me
the required archive logs
now for testing purpose we need to get the DB up at our site but the archive
logs are not there, the client will send the same but it will take time
Is there any way to get the DB up without
Title: RE: Large Export Problem ..
Yes, that was a typographical error in my response. I meant IGNORE=N.
-Original Message-
From: Janardhana Babu Donga [mailto:[EMAIL PROTECTED]]
Jacques,
Your response is helpful. Thanks for your suggestions. I think I should import with
I've taken it and can't say enough good things about it. I haven't taken any
other course from which I gained as much practical and useful knowledge. If
you can attend one training this year (particularly if you're looking for an
Oracle performance/optimization course), go to this oneI
It looks to me as if you've worked out exactly
what the parameter does already.
Under OPS, (and RAC) each instance has its own
SCN range, and the ranges are not re-synchronised
in real time (in general). SCNs are resynchronised
by piggybacking when messages pass from instance to
instance.
Hi
Vladmir,
You
could find it from - select instance_number from v$instance;
You
can also query gv$instance to find all sessions connected to both the instances.
This query can be run from any instance and it will display all the sessions
connected to all the instances. There's a
Jeremiah,
The only reason I could forsee is if the split is not atomic. This is the
case
in our (Non-SAN) environment. The split takes several minutes and the
control
files have been mis-matched when the DB copy was started on the reporting
server.
My solution under V8.0 was to use only one of
Babu,
It may not be THE correct answer but I keep all of the packages
created by the development community on a disk as the txt or sql
statemants used to create them. It helps to have them handy when they
need to be modified because of a table structure change, etc. I just
give them back to the
alter database open resetlogs;
you won't be able to apply the archived logs he sends over later on
(yes Jeremiah I know there is a way to recover through a resetlogs but
they will be doing work against the database in the interim!)
but if data loss is not a problem, that's the way to go
Rachel
-Original Message-
Sent: Wednesday, March 26, 2003 11:54 AM
To: Multiple recipients of list ORACLE-L
I've taken it and can't say enough good things about it. I haven't taken
any
other course from which I gained as much practical and useful knowledge. If
you can attend one training
I go even one step further here. Each object is required to have it's own creation
script. A
package will have two scripts, one for the header and one for the body. Also the
script name has
the package revision included in it as well. For example:
hyp_packages.115.00.sql
I haven't taken it yet, but I am signed up for the Chicago Clinic on April
8-10Can't Wait! I've heard really good things about it.
At 08:04 AM 3/26/03 -0800, you wrote:
Anybody take this course before. Any comments, good or bad. Thanks
--
Please see the official ORACLE-L FAQ:
Raj
I haven't tried reading logs without Logminer, but I have used Logminer
without the dictionary. It is tough. Basically, everything is in hex. I
would guess the who is the userid in hex. You could list the usernames and
user_ids from dba_users and convert the user_id column to hex and see
Well thanx a lot, Rachel and Jose.
At 12 midnight the mind seems to have stopped working
:-)
Regards
Naveen
-Original Message-
Sent: Thursday, March 27, 2003 12:21 AM
To: Multiple recipients of list ORACLE-L
alter database open resetlogs;
you won't be able to apply the archived
Title: RE: alter system dump logfile ... question
Thanks Dennis,
That's what I am trying to find ... which column to match ... and yes in past I have used logminer with a lot of success ...
Raj
-
Rajendra dot Jamadagni at espn dot
we are test the XNL to ORACLE load function. we
export data from Crystal
Report as XML format. on ORACLE 9ir2, we used
oraxsl utility tried to
convert it from XML== plain text (we can use
SQL/Loader load into ORACLE
DB).
We can NOT find any xmltotext.xsl file (style sheet)
under ORACLE
9ir2.
Joe, have you tried the Veritas support site?
There's a lot there, sometimes too much.
Jared
Testa, Joe [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/26/2003 10:04 AM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Jonathan Lewis wrote:
However, it is possible that you managed to
change the number of db block buffers available
quite dramatically - which could have an effect
on the size at which a table is considered to be
a 'small' table, which could have an impact on
the actual resource usage for the
Roland,
It took me exactly 12 seconds to find out exactly how to do this
from the documentation. This was simply by clicking on 'CREATE SEQUENCE'
in the SQL Reference manual.
I strongly suggest you learn to use and read the documentation.
Jared
[EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
Title: RE: events number and meaning
-Original Message-
From: Deshpande, Kirti [mailto:[EMAIL PROTECTED]]
...
Also check www.oraperf.com, there a list of wait events by Oracle Versions.
Free membership is required.
I registered and logged in. When I click on wait events on the
Ron,
It seems to be the correct answer I believe. If the package is lost I should
ask the developer to get it from the version control system and I should
compile it. This seems to be the right approach rather than importing the
dump file and get a big list of errors.
After seeing the responses
Hello,
Env: Oracle 9.2.0.2.0 on Solaris 9 (2 machines)
I'm trying to set up a daily cloning process between 2 Oracle
instances (SIDA is source, SIDB is target) using the CREATE CONTROLFILE
REUSE SET DATABASE SIDB ... method. The steps are:
1) Shutdown immediate SIDB
2) Shutdown immediate
I'm trying to add a foreign key constraint to a table, and specify the
tablespace and pctfree. Here is the syntax I'm using:
alter table jobproductoutrigger add (
constraint fk_jobprodu_currjobfa_currjobf foreign key (
jobnbr, sourcefiscalyear )
references currjobfact (jobnbr,
I like RCS for this. Any version you need can be retrieved.
If a large number of packages are used for an app, you could
take it a step further and use CVS. This would allow app level
releases as well.
Jared
Ron Thomas [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/26/2003 11:09 AM
Title: dropped sequence - backup/recovery question
The question from Janardhana Babu Donga yesterday about how to recreate a dropped package/procedure/function made me think about something else. Imagine the following scenario:
There is some commercial off-the-shelf (COTS) software package
I've taken it twice (one of the first public classes and again the when they
started offering the reduced alumni rate), and stand by my quote on Hotsos's
website.
If you have the time to take a class, take the Hotsos Clinic. It was the
absolute best performance tuning resource that I have ever
Poking around metalink I found a reference to a utility called
sysresv (note 123322.1). According to the document it
displays the id and key for each for the shared memory segments
which the Oracle instances have created.
I tried it on an AIX 4.3 system running three Oracle instances and it
Dennis,
I don't think you can add an index this way. I've always had to add the
index separately. It would be great if this was possible.
Tom Mercadante
Oracle Certified Professional
-Original Message-
Sent: Wednesday, March 26, 2003 3:04 PM
To: Multiple recipients of list ORACLE-L
Dennis...
I'm trying to add a foreign key constraint to a table, and specify the
tablespace and pctfree. Here is the syntax I'm using:
alter table jobproductoutrigger add (
constraint fk_jobprodu_currjobfa_currjobf foreign key (
jobnbr, sourcefiscalyear )
references currjobfact
Bill
If a tablespace is read-only, it should recover fine, nothing to write.
I don't think you can recover a tablespace from another database. I think
they will have different SCN numbers. I think only read-only tablespaces are
allowed with different SCN numbers. However, before backup, you
Jared, You have entirely to much free time if you are timing each of
your work items. OR is it the compulsive disorder that DBA acquire?
RON 1.2.3.4.5.6.7.8.
[EMAIL PROTECTED] 03/26/03 02:28PM
Roland,
It took me exactly 12 seconds to find out exactly how to do this
from the documentation.
Dennis,
I don't think, you can use using index clause, when adding FOREIGN KEY
constraint, you can use it though with PRIMARY KEY and UNIQUE constraints.
For FOREIGN KEY constraint you should create index with separate DDL.
Probably, because foreign key can exist without index (though it's not
.
University officials have known about the problem for years and recently told
state auditors they had no plans to fix the glitch, saying it would be too
costly and time-consuming. Instead, the CSU said it decided to require
employees to sign confidentiality agreements, promising to
you might also want to capture the database outline in OEM (in change
manager)
then if you happen to lose a procedure - u can ask oem to do a diff between
yr database and the stored outline and then ask it to generate a script to
sync them.
babu
- Original Message -
To: Multiple
Title: RE: Syntax for foreign key
Indexes are only automatically created to enforce primary key and unique constraints.
If you want an index to match the columns of a foreign key constraint, you will have to do it separately.
P.S. If you already have a non-unique index on the PK or unique
It was 11.7 seconds, but I rounded up.
Jared
Ron Rogers [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/26/2003 12:33 PM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Subject:Re: What is wrong with this?
FWIR, CVS doesn't handle binaries tho. We do the same thing for rdf's (report files)
and fmb's
(forms files) too.
Ron Thomas
Hypercom, Inc
[EMAIL PROTECTED]
Each new user of a new system uncovers a new class of bugs. -- Kernighan
Login as the owner of the shared mem segments and run it.
It just worked for me on RH 7.2 with 8.1.7
Jared
Schauss, Peter [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/26/2003 12:33 PM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Jacques,
If you know the field that is being populated by the sequence, examine
the last 25/50 entries, determine the pattern and 'high' value, then recreate
the sequence accordingly.
If you don't know the field, you can take a guess that it is the primary
key (or perhaps another unique
Title: RE: Database Modeling- Normalization - Dinosaurs or What?
Yes. Also, I don't. Please tell me why that matters - I think you are onto something here.
Oracle OCP DBA
-Original Message-
From: DENNIS WILLIAMS [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, March 25, 2003 2:39 PM
To:
Title: RE: Veritas Quick I/O for Oracle
Dennis,
The benefits of QIO are realized not only by Oracle but also by the storage administrators. Oracle improvements come from kernelized async I/O, elimination of UNIX double-buffering and single-writer file header locking in the O/S. Storage
I've not bothered to try this myself, but I think you
will need to dump the first redo record of the transaction
to get the 'transaction audit' change vector that is the
last vector of the first record.
In Oracle 8.1, this would give you the userid of the
transaction owner - but this looks like
The link has been fixed.
Thanks to Anjo for his quick response.
Jared
Jacques Kilchoer [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/26/2003 11:28 AM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Subject:
Thanks for the update Jonathan. The recommendation was made by Oracle
Support, after they noticed a lot of on going read of SCN to complete
waits from a system dump. Almost 40% of the waits. In some of my statspack
report, I even see this waits accouting for 70% of the total. Setting this
This looks symptomatic of a problem with
using a temporary table in a join. Because
of anomalies with statistics on temporary
tables, Oracle has chosen a completely mad
execution plan - e.g. involving the temporary
table as the inner table of a nested loops join.
(Not guaranteed, but reasonably
Hi,
I have to recover my database back to 23-MAR-03.There was some changes to the data
yesterday and cannot be reverted.
We want to recover the data as of 23-MAR-03.Database is running on archive log mode
and I have the hot backup datafiles
of Feb 3 and all the archive logs since then till
1 - 100 of 126 matches
Mail list logo