--
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 http://www.fatcity.com
San Diego, California-- Mailing list and web hosting services
Hi Peter,
pooling is for:
user that connected but idle will be disconnected temporary for new incoming or
outgoing request.
ON or BOTH or YES or TRUE is to turn on
OFF or FALSE or NO is default to turn off
IN for incoming connection only
OUT for outgoing connection only
Sinardy
Hi,
Restartup your database if possible.
If you set the datafile autoextent on then you have to recreate that tablespace,
if you set it to not autoextent on for some query may fail due to not enough sorting
space.
Sinardy
-Original Message-
Sent: 12 November 2003 14:24
To: Multiple
Hi Mladen,
Don't desert me now, I thought we've come a long way !!
Although, I'm not a listed Oracle guru, let me try and explain further ;)
PCTFREE works in almost exactly the same way as it does for non-ASSM
objects. PCTFREE determines how much of the block we want to reserve for
subsequent
Many thanks. I am back on again.
Regards
Lee
-Original Message-
Sent: 11 November 2003 21:59
To: Multiple recipients of list ORACLE-L
-Original Message-
From: Robertson Lee - lerobe [mailto:[EMAIL PROTECTED]
Sent: Wednesday, 12 November 2003 00:45
To: Multiple recipients
Yep, the situation can get bad for parallel execution, especially if blocks
read aren't cached...
But for serial FTS I haven't seen such a problem, I did even a test to
verify it on 9.2.0.4, and did see behaviour as I expected - all blocks were
scanned using multiblock reads and rows were returned
List, i need some help on this... i am trying to get information on this but
, and the patches available,
the bug no is 1809113. but metaling does not allow to display this bug !!!
even when i use my PARTNERS
login !!!
a.. The bug is not classified as publicly accessible (non-public).
b..
Rahul
Are you certain you have the correct number? What is the bug about -- you
may be able to search using alternate words.
Dennis Williams
DBA
Lifetouch, Inc.
[EMAIL PROTECTED]
-Original Message-
Sent: Wednesday, November 12, 2003 6:30 AM
To: Multiple recipients of list ORACLE-L
Hmmm ... you saw only a through d ... that's why ...
There is another bullet point ..
e. This bug is caused by one of our stupid developers, but it is embarrassing for us
to discuss that in public.
ps: if you want details on that bug, log and itar ... and pray.
Raj
Hi
It is posible to create other schemas on a logical stand by database ?
I mean, schemas that don?t exist in the primary database.
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Juan Miranda
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051
yes. Well documented in the manuals
--- Juan Miranda [EMAIL PROTECTED] wrote:
Hi
It is posible to create other schemas on a logical stand by database
?
I mean, schemas that don?t exist in the primary database.
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
I don't usually forward my reader email to the list, but the
question below strikes me as rather interesting. In this
case, SQL*Loader appears to be causing all SQL statements
that refer to the table being loaded to be invalidated. Is
this normal behavior? Does anyone know why it might be the
Hi
We do something similiar, but instead of deleting the tables beforehand, I
just use the SQL*LOADER REPLACE option. No such problems as described in the
original eMail occured so far. The platform is Oracle 9.2.0.3 on Win3k.
Regards,
Stefan
-Ursprüngliche Nachricht-
Von: Jonathan
Windows 3000?
-Original Message-
Sent: Wednesday, November 12, 2003 7:44 AM
To: Multiple recipients of list ORACLE-L
Hi
We do something similiar, but instead of deleting the tables beforehand, I just use
the SQL*LOADER REPLACE option. No such problems as described in the original
Or you can
1. Create another Temporary Tablespace with CREATE TEMPORARY TABLESPACE
new_temporary_tablespace
TEMPFILE . SIZE ... AUTOEXTEND ON NEXT .. MAXSIZE ...
2. Issue ALTER USER TEMPORARY TABLESPACE new_temporary_tablespace commands
3. If running 9.2 with a Default Temporary Tablespace,
Im reading the Sybex OCP book on tuning and it is absolutely loaded with inaccuracies.
Is the test the same way? If so do they improve it in 9i?
The book is loaded with all types of hit ratios, discussions about committing
frequently to IMPROVE performance, and other garbage.
anyone know the
Juan,
How can you create Schema in STANDBY database that does not exist in Primiary
database?? Can you give me an example?
Muqthar Ahmed
-Original Message-
Sent: Wednesday, November 12, 2003 8:24 AM
To: Multiple recipients of list ORACLE-L
yes. Well documented in the manuals
I never use those. I find it better not to be overly smart with sqlplus.
Sqlplus is a program for ad-hoc queries from the command line and for
the administrative functions (startup/shutdown, alter database/system),
nothing more and nothing less. If I need a program, I can either write a
PL/SQL
Short answer: look in the manuals.
Longer answer: a LOGICAL standby database does not get updated in the
same manner as a PHYSICAL standby database. Logical Standby is simply a
database which is updated via SQL statements generated from the
archived redo logs, not application of the archived redo
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using ASO,
and third party authentication (securID)
we have configured sucureID authentication for our DB, but out back office
app using forms6i does not connect to the DB when this is
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using ASO,
and third party authentication (securID)
we have configured sucureID authentication for our DB, but out back office
app using forms6i does not connect to the DB when this is
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using ASO,
and third party authentication (securID)
we have configured sucureID authentication for our DB, but out back office
app using forms6i does not connect to the DB when this is
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using ASO,
and third party authentication (securID)
we have configured sucureID authentication for our DB, but out back office
app using forms6i does not connect to the DB when this is
I believe it's direct read from files in parallel execution, nothing gets
read from cache.
Cached blocks for the table get flushed to files before the direct read.
Regards,
Waleed
-Original Message-
Sent: Wednesday, November 12, 2003 7:10 AM
To: Multiple recipients of list ORACLE-L
I wonder whether the invalidation comes about from the use
of TRUNCATE, which is considered a DDL statement. I'd guess
that any DDL to a table would invalidate existing SQL
statements.
Best regards,
Jonathan Gennick --- Brighten the corner where you are
http://Gennick.com * 906.387.1698 *
The question was about LOGICAL standby database.
From your point of view it is like a regular database and you can create
schema in it.
Yechiel Adar
Mehish
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Wednesday, November 12, 2003 4:14 PM
Juan,
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using ASO,
and third party authentication (securID)
we have configured sucureID authentication for our DB, but out back office
app using forms6i does not connect to the DB when this is
I'll try tracing the session.
Global.client_dim is just a table with client info and a column that
corresponds to client logins to enable row level security.
Thanks for the suggestions.
As stated earlier..i'll post the resolution.
chris
-Original Message-
Sent: Tuesday, November 11,
alright already. I think we got the message !
;-)
-Original Message-
Sent: 12 November 2003 14:30
To: Multiple recipients of list ORACLE-L
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using ASO,
and third party
I never use those. I find it better not to be overly smart with sqlplus.
Sqlplus is a program for ad-hoc queries from the command line and for
the administrative functions (startup/shutdown, alter database/system),
nothing more and nothing less. If I need a program, I can either write a
PL/SQL
That's the target year for the stable and secure release of MS Windows.
HAL:What are you doing Dave? David Bowman: I'm turning you off, HAL.
On 11/12/2003 08:49:26 AM, Nelson, Allan wrote:
Windows 3000?
-Original Message-
Sent: Wednesday, November 12, 2003 7:44 AM
To: Multiple
IIRC, direct path loads invalidate indexes. As the execution plan may use this
index, it makes sense that any statements referencing these tables will be invalidated
and need to be
reparsed. I've never tested this, but it makes sense (at least Oracle sense).
As for the 4031 errors, could they
sandali- may tatapusin lang ako. I'll call you in 30 minutes.
wrote:
You're much too nice.
Rachel
Carmichael To: Multiple
recipients of list ORACLE-L [EMAIL PROTECTED]
wisernet100 cc:
As a strange thing, from 10046 trace I saw that normal table scanning was
done using direct reads, this was expected behaviour, but the lookups of
migrated rows were reflected as 'db file sequential reads'. And even more,
there were 3 subsequent sequential read waits for the same datablock in a
Enough already!
Igor Neyman, OCP DBA
[EMAIL PROTECTED]
-Original Message-
rahul
Sent: Wednesday, November 12, 2003 9:30 AM
To: Multiple recipients of list ORACLE-L
the bug no. is correct, it relates to
forms 6i (running on web) is not able to connect to the DB when using
ASO,
and
Perhaps you could clarify what you mean by
'it grows up faster and faster'.
What exactly does that mean, what script or tool is
showing you that, what is its output, and why is
it a problem?
Restarting the database seems extreme advice for
such a meager description of a problem.
Jared
On Tue,
Hmm...
I'd like to know where in the manuals... :-)
I do not think so since the standby database stay in
permanent recovery mode.
JL
--- Rachel Carmichael [EMAIL PROTECTED] wrote:
yes. Well documented in the manuals
--- Juan Miranda [EMAIL PROTECTED] wrote:
Hi
It is
Tanel,
Maybe a PQ FTS needs to resolve migrated rows immediately as it's possible
that the migrated row is located in a block that's allocated to a different
PQ slave.
Chris
-Original Message-
Sent: 12 November 2003 15:49
To: Multiple recipients of list ORACLE-L
As a strange thing,
Ryan - I took the 8i OCP for Tuning. I used Couchman to study with, and
don't recall any big differences with the exam. The exam is prepared from
the Oracle Education Student Guide for Oracle 8i. That would have been a
couple of years ago, and if Oracle Education was teaching hit ratios back
then,
Hi,
I'm currently studying for this exam but can't find info to say whether the
exam covers 9.2 or just 9.0. Anyone any clues
Thanks,
Chris Dunscombe
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Dunscombe, Chris
INET: [EMAIL PROTECTED]
Fat City Network
Using PQ, the segment is split into multiple sub-segments using rowid range
scan.
Since there is no guarantee that the migrated row will be in the same
sub-segment that has the pointer to the migrated row, the PQ slave might
need to resolve the issue real time.
I mean the migrated row might
There are two kinds of standby database: physical and logical. The
original post referred to logical standby. The normal state of a
logical standby database is open, and it can contain schemas and
objects that do not exist in the primary.
Check the manuals again! ;-)
--- Jose Luis Delgado
I think you're confusing physical and logical standby databases.
Logical standby databases aren't in recovery mode.
--Walt Weaver
Bozeman, Montana
On Wed, 2003-11-12 at 09:09, Jose Luis Delgado wrote:
Hmm...
I'd like to know where in the manuals... :-)
I do not think so since the
Is there a primary key on the table ?
Regards,
Denny
Jonathan Gennick wrote:
I don't usually forward my reader email to the list, but the
question below strikes me as rather interesting. In this
case, SQL*Loader appears to be causing all SQL statements
that refer to the table being loaded to be
Jose Luis,
What you say refers to the physical standby database (which works well), not to the
logical standby database (which on the paper looks great, allows you to open the
database, create additional tablespaces, create additional indexes on replicated
objects etc) but which in practice
Runs on 9.2 but there's nothing specific to 9.2 rather than 9.0, IIRC.
Pete
Controlling developers is like herding cats.
Kevin Loney, Oracle DBA Handbook
Oh no, it's not. It's much harder than that!
Bruce Pihlamae, long term Oracle DBA.
-Original Message-
Dunscombe, Chris
Sent:
Maybe, the process that read the block use the same logic to pass the rows
out, no matter if it FTS or direct block read. So this logic resolved
migrated rows in the block that was passed to him.
Yechiel Adar
Mehish
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL
took me exactly two clicks to find the following:
Oracle9i Release 2 (9.2) New Features in Data Guard
The features and enhancements described in this section were added to
Data Guard in Oracle9i release 2 (9.2).
* Logical standby database
Until now, there has been only the physical
Jonathan,
I don't see where the TRUNCATE command is used in the original request
message.
The 20 staging tables have a delete action against them.
The data is sqlldr action to load the tables.
The final tables are loaded from the staging tables using a PL/SQL
procedure.
The final tables are
Jonathan,
I don't see where truncate is being invoked, though I am not a big user of sql*loader.
Truncate should invalidate statements.
Daniel
Jonathan Gennick wrote:
I wonder whether the invalidation comes about from the use
of TRUNCATE, which is considered a DDL statement. I'd guess
Stephane,
What sort of problems can one expect from logical standby?
I'm toying with the idea of using it as a replication database -- no
additional schema objects will be created, but users will have read-only
access to it. It's one of the options I'm looking at.
Seems to me like there was a
Chris
I'm betting on 9.0. For it to cover 9.2 would have meant that Oracle
would have had to go back and recreate the test. And Oracle would have felt
compelled to change the name of the test. However, I think it possible that
any question whose answer would be true for 9.0 but false for 9.2
There are two options to replace all data in the table: REPLACE and TRUNCATE
which are equivalent to truncate and delete sql statements. If you have
staging tables without RI or triggers then use truncate. Using delete just
takes a lot longer and use a lot more resources.
We use TRUNCATE almost
Hi all,
As Stephane told, logical standby (LSB) has a lot of tiny little
exceptions and special issues to cope with. I've done one implementation
in production until now (did appr. 20 Physical Standby sites as well).
But, even that site uses 2 LSB's as reporting systems, and has a PSB for
the
yeah but...
if you attempt (as I do) to isolate the schema owner from the users
which have select/insert/update/delete privileges, TRUNCATE won't work
unless you have granted DROP ANY TABLE (I *really* hate that that is
required) to the account which does the actual sql load. So I use
REPLACE,
there was a thread -- Paul Baumgartel started it looking for
information on logical standby.
IIRC, he found that there were a few gotchas -- check the fatcity
archives.
I do know that since it's based on Logminer technology, it has the same
limitations that Logminer does
--- Walt Weaver [EMAIL
Title: New interface on Metalink to do queries ...
I found it today ...
http://metalink.oracle.com/metalink/plsql/ml2_documents.showFrameDocument?p_database_id=NOT_id=102421.1
Raj
Rajendra dot Jamadagni at
Daniel,
How does using the TRUNCATE command is a sqlldr invalidate anything?
The sqlldr truncate command reuses the storage that the table originally
used and does not change the HW mark. If there are indexes on the tables
then they are placed in the DIRECT PATH state during the load and
updated
KamYee,
Can we take a look at the SQL*Loader control file? Also the output of
select * from v$sql where lower(sql_text) like '%yourstagingtable%' and
invalidations 0
Yong Huang
--- Jonathan Gennick [EMAIL PROTECTED] wrote:
I wonder whether the invalidation comes about from the use
of
And yet it STILL returns articles about DEC's RDB.
Wasn't the new/improved MetaLink supposed to be out Q1 of *this* year? All
I want is valid HTML, no %$#*( frames, and no hits that I can't view or that
don't apply to my search (e.g. DEC RDB, Apps, etc.).
Rich
Rich Jesse
Hi.
I'm running tkprof on my PC with a version 9i client.
The trace file was generated on a Solars version
8.1.7.4 database. (I'm using tkprof on 9i to get wait
statistics.) The trace is 10046 level 12
I assume the time waited is in addition to the elapsed
time for the call -- correct?
Since
Ron,
I don't know about the TRUNCATE option w/ sql*loader, but the regular DDL
TRUNCATE invalidates sql that references the table.
Example:
1 select sql_text, invalidations
2 from v$sql
3* where sql_text = 'select * from emp'
SQL /
SQL_TEXT
im going to take it soon. I was going to just read howard rogers guide then the otn
one.
you think that is enough? I just want to pass it and get my piece of paper. I already
know the 9i stuff that is useful to me.
From: DENNIS WILLIAMS [EMAIL PROTECTED]
Date: 2003/11/12 Wed PM 12:19:32
My company started offshoring jobs to India. Ive been told Im safe, but in a corporate
environment, your boss doesnt know much. He didnt even get advance notice about the
layoffs that occurrted today. I believe long term they want to outsource the whole
department overseas. My boss doesnt know.
On 11/12/2003 12:59:25 PM, Jesse, Rich wrote:
And yet it STILL returns articles about DEC's RDB.
Wasn't the new/improved MetaLink supposed to be out Q1 of *this* year? All
I want is valid HTML, no %$#*( frames, and no hits that I can't view or that
don't apply to my search (e.g. DEC RDB,
Yep, truncate invalidates parsed SQL:
SQL select count(*) from
t3;
COUNT(*)--
0
SQL select sql_text, optimizer_mode,
invalidations from v$sql where sql_text like '%select count(*) from
t3%'and sql_text not like '%v$sql%';
SQL_TEXT
OPTIMIZER_
Or... Create a stored procedure that truncates the table... Grant execute
on the procedure to the user... The user executes the procedure and then
calls sqlldr...
Tim
-Original Message-
Sent: Wednesday, November 12, 2003 12:45 PM
To: Multiple recipients of list ORACLE-L
yeah but...
Walt,
I'll step in here--my experiment with Logical Standby convinced me that
it is not ready for prime time.
1. Major bugs caused apply process to crash repeatedly.
2. Difficulty filtering out DDL from apply stream.
3. Horrendous performance of apply process--frequently the elapsed
time to
Barbs, what this trace file tells you is that the client program
takes much more time delivering the data to oracle then it takes
for oracle to deliver the data to the program. In other words, your
program takes its time between the calls to the database. Maybe you
should run the profiler and see
Barb,
This is a great example of where reading the trace file will tell you
what you need to know. Is the 10 centiseconds of time in 8 1.25
centisecond events or in 1 10 centisecond event and 7 .1 centisecond
events? The location (parse/execute/fetch/post tx) are also important.
Daniel
Barbara
I'm studying for the 9i Performance Tuning exam, too. I'm glad to hear about
the inaccuracies in this book. I have this book and the Oracle Press book by
Pack. I also have Oracle Online Learning and I think I will stick more
closely to that. Also, hopefully the Self Test Software gives a good
Walt, drop me your email-address, and I send you the handouts of a
special I presented about DG for Oracle University in
Stockholm.
I'm going out now for a few hours (it's 19.30 over here), but I'll
respond later this evening.
regards, Carel-Jan
At 09:19 12-11-03 -0800, you wrote:
Stephane,
Barb,
I think tkprof prints output in seconds, so I think you're looking at
10.73 seconds of response time here.
First, some definitions:
- The SQL*Net message from client event maps to an OS read() call to
the file descriptor to which SQL*Net is connected.
- The SQL*Net message to client
Walt Weaver wrote:
Stephane,
What sort of problems can one expect from logical standby?
I'm toying with the idea of using it as a replication database -- no
additional schema objects will be created, but users will have read-only
access to it. It's one of the options I'm looking at.
In other words, you can finish performance analysis of the client
only by finding out where the time is spent on the client side.
If it's an interactive program and you have a homo sapiens interacting
with it, then think time is a factor. If, on the other hand, it's
a batch program, then run a
Hey all,
After reading the article in the latest OraMag about 9i's CDC -- Change
Data Capture and not Centers for Disease Control (cool reading that MMWR
is) -- I'm wondering how most folk get that CDC data to the staging area of
a DW/DM. The three ways I can think of are by using DB links,
Wednesday, November 12, 2003, 12:04:35 PM, Daniel Fink ([EMAIL PROTECTED]) wrote:
DF I don't see where truncate is being invoked, though I am not a big user of
sql*loader.
DF Truncate should invalidate statements.
Sorry. He mentioned in a later email to me that he was using
the TRUNCATE option
Daniel, it's seconds, not centiseconds. The tkprof writes things down
in seconds. If it's a 9i trace, then the times within trc files are in
microseconds. In trc file produced by 8i, it's centiseconds. By some magic,
tkprof knows the difference and usually gets the right times.
On 11/12/2003
Daniel:
Does this mean I was supposed to be paying attention
in class??
Here's everything I can find associated with sql*net
message from client in that cursor. Can't figure out
from this how it came up with 10.73 (although that one
wait matches up with the max wait of 10.71)
Am I being
We are being asked by Auditing to stop using the SYS, and SYSTEM
accounts. They would like for us to create an Oracle Role with the same
permissions a SYS and SYSTEM, then grant the role to each of the DBA's.
Don't ask me why. Nothing is being audited in 99% of the databases.
They just say it in
Wednesday, November 12, 2003, 12:59:25 PM, Jesse, Rich ([EMAIL PROTECTED]) wrote:
JR And yet it STILL returns articles about DEC's RDB.
But isn't that the solution to all Oracle problems? Just
upgrade to DEC RDB?
Best regards,
Jonathan Gennick --- Brighten the corner where you are
Daniel,
I understand what is you are saying and what you tested but I don't
the why or what it means. Does it mean that the sql command is not
going to work? Does it means that you have to issue it again to get it
to work?
Ref:
Doc ID: Note:123214.1
invalid
Type: PROBLEM
Status:
That won't work if you're using RMAN.
The account that makes the backup needs to be able to do so as sysdba.
You can't grant that through a role.
The reason for separate accounts is accountability. But if you're not auditing,
that won't help much, as you already stated.
Jared
Smith,
There is a difference between a statement that is invalid (i.e. won't
execute) and one whose execution plan has been invalidated in the shared
pool. When the statement is 'invalidated' all it means is that if a
process wants to reexecute the statement, it must be reparsed. At this
time, the
Thanks for the info on logical standbys everyone.
The opinions on it seem rather unanimous.:)
--Walt
On Wed, 2003-11-12 at 12:14, Stephane Faroult wrote:
Walt Weaver wrote:
Stephane,
What sort of problems can one expect from logical standby?
I'm toying with the idea of using
Barbara,
I know how distracted you were with all the jokes being tossed about,
but really...
It looks to me (with just this snippet of trace) that the cursor #3 did
some communication and then waited for 10 seconds for a response.
Without seeing the raw trace file and sequence of events, this is
Or if you're auditing in a pre-9i DB, which won't audit SYS and SYSDBA.
Rich
Rich Jesse System/Database Administrator
[EMAIL PROTECTED] Quad/Tech Inc, Sussex, WI USA
-Original Message-
Sent: Wednesday, November 12, 2003 2:30 PM
To: Multiple
Personal Opinion here: I don't use SYS or system for anything where it is not
absolutely required. All of the DBA's have the DBA role granted to them we log on
as ourselves. This is simply so that we don't accidentally step on something really
important. In general one should never create
Title: RE: Stop using SYS, SYSTEM?
Jared,
I followed Robert Freeman's advice and created an RMAN user in all my DBs called backup_admin with SYSDBA privilege so that RMAN doesn't use SYS or SYSTEM. This allows you to change system passwords at will and not interfere with backups. Works just
I agree 100% with Dick. Nobody should be using SYS or SYSTEM. If RMAN
requires a SYS connection, then so be it.
Tom Mercadante
Oracle Certified Professional
-Original Message-
Sent: Wednesday, November 12, 2003 3:45 PM
To: Multiple recipients of list ORACLE-L
Personal Opinion here:
Smith, Ron L. scribbled on the wall in glitter crayon:
We are being asked by Auditing to stop using the SYS, and SYSTEM
accounts. They would like for us to create an Oracle Role with the
same permissions a SYS and SYSTEM, then grant the role to each of the
DBA's. Don't ask me why. Nothing
Daniel,
That is what I thought I read in the doc's,
Thanks,
Ron
[EMAIL PROTECTED] 11/12/2003 3:29:27 PM
There is a difference between a statement that is invalid (i.e. won't
execute) and one whose execution plan has been invalidated in the
shared
pool. When the statement is 'invalidated' all
We avoid using SYS as much as we can, but we use SYSTEM ... cautiously I might add.
Raj
Rajendra dot Jamadagni at nospamespn dot com
All Views expressed in this email are strictly personal.
QOTD: Any clod can have
David
You can remove the create session priv from the RMAN user and this
will make a little
harder for most users to connect, but RMAN will work fine :-)
David Wagoner wrote:
Jared,
I followed Robert Freeman's advice and created an RMAN user in all my
DBs called backup_admin with SYSDBA
Ron,
It is a good practice, in general, to stop using SYS and SYSTEM accounts for
everyday use. The simplest rule of thumb is accountability somehow increases
many times over when you link a database named user to a physical person,
not a ethereal entity like SYS. This is especially true if you
Smith, Ron L. wrote:
We are being asked by Auditing to stop using the SYS, and SYSTEM
accounts. They would like for us to create an Oracle Role with the same
permissions a SYS and SYSTEM, then grant the role to each of the DBA's.
Don't ask me why. Nothing is being audited in 99% of the
And for an opposing opinion:
Let's see now. We create another user and grant that user all the
privileges needed to do ANYTHING. And that makes things so much more
secure? If that's the prevailing thought among the database world, then
it's safe to say that the Unix admins have infinitely
And as Arup's Oracle Magazine's DBA of the Year for 2003, he's probably
right.
Congrats, Arup!
Rich
Rich Jesse System/Database Administrator
[EMAIL PROTECTED] Quad/Tech Inc, Sussex, WI USA
-Original Message-
From: Arup Nanda [mailto:[EMAIL
Barb,
What you're really after is called forward attribution in the book.
You need to see the dbcall that immediately follows each WAIT
nam='SQL*Net message from client' event in the trace file. This will
tell you what database call's execution it was that ended the read().
Cary Millsap
Hotsos
Yep, that's exactly what I thought, that due migration the row might go to a
PX granule read by another slave... but I still don't see enough reason why
migrated rows are handled different from serial scan. During serial scan you
also might have rows in beginning of segment migrating to end of it,
1 - 100 of 125 matches
Mail list logo
