Don
We have successfully used forms 6i against Oracle 9i databases (9iR1
and 9iR2).
Just a note that 9iAS Rel1 comes with forms6i.
Only in 9iAS rel 2 and above are forms 9i used.
Just make sure the correct forms 6i patchset has been applied that
fixes the
relevant forms bugs.
Kind Regards
At 18:24 17-11-03 -0800, you wrote:
Hi,
Since everyone is just guessing, can you say more detail about the
bad performance? High CPU usage with no idle so client wait for CPU time?
or Low CPu usage but client still wait? or High disk io rate with 100%
disk busy?
Can you show us some
Bambi,
in earlier versions of unix, it would reply don't know how to make love like ...
[EMAIL PROTECTED] uname -a
dgux orion R4.20MU06 generic AViiON PentiumPro
[EMAIL PROTECTED] make love
Make: Don't know how to make love. Stop.
but now ...
[EMAIL PROTECTED] uname -a
AIX ariel 1 5
Your developer is wrong ... dead wrong. We use 6i in C/S mode against 9012/9202 since
about Sept'02.
Raj
Rajendra dot Jamadagni at nospamespn dot com
All Views expressed in this email are strictly personal.
QOTD:
Title: Message
We are
working a SharePlex project from Quest. For our purposes (Financials) with
heavy adhoc (Cognos) we need to stay as close as possible to the production
instance. The advantage to SharePlex is that it is a log miner and hence
does not do a lot of sql against the source
Peter,
Do you mean with starting the database? Not at all. The db will register
when the listener comes up.
Tom Mercadante
Oracle Certified Professional
-Original Message-
Sent: Monday, November 17, 2003 4:04 PM
To: Multiple recipients of list ORACLE-L
If you remove the database
Hello all,
I'm attempting to import into 9.2.0.1.0 from 7.3.4 and I'm getting loads of
the same error:
IMP-00017: following statement failed with ORACLE error 9275:
GRANT SELECT ON DOWNLOAD_SEQ TO INTERNAL
IMP-3: ORACLE error 9275 encountered
ORA-09275: CONNECT INTERNAL is not a valid DBA
Barry,
Why make life difficult? It's just a role, not a data
object referenced by applications (hopefully). Change it's
name to something that is not a reserved word and move on.
There is a list of reserved words in the SQL Language
reference.
Hope this helps...
-Tim
Hello all,
I'm
I believe you license based on the 'actual' number of cpu's, BUT you can
only license standard edition on a machine that has a maximum CPU count of 4
or less. Maybe that is what Jared is remembering.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent:
Tim,
I checked v$reserved_words. At least in 9.2.0.1, INTERNAL is not in there.
Oracle should address this issue.
When I trace the SQL GRANT SELECT ON SOMETABLE TO INTERNAL, it stops in
parsing.
Nonetheless, it's confusing to say the least to create a role called internal.
Yong Huang
--- Tim
Oops! blush Thanks for the correction, Yong!
Tim,
I checked v$reserved_words. At least in 9.2.0.1, INTERNAL
is not in there. Oracle should address this issue.
When I trace the SQL GRANT SELECT ON SOMETABLE TO
INTERNAL, it stops in parsing.
Nonetheless, it's confusing to say the
Yong,
It seems to me that you are missing the point here. The real point is that
you should not have granted select on some table to internal - ever. And
a new release caught you on this mistake. And now, you have to fix it. It
is not an Oracle problem, but a mis-use of an Oracle internal
As of yet I'm unsure how the application would be affected if I rename the
role - I need to do some investigation.
I tried this in Oracle 8 and it worked fine - It just seems to be oracle 9
that doesn't like it.
I've also checked v$reserved_words and INTERNAL is not listed, INTERNAL_USE
and
This would explain why it worked when I tried it on oracle 8 then.
The developers that originally created the application left quite a while
ago, so I don't think I'll be able to ask them why they did it this way.
But basically you're saying that it shouldn't have been done like this and
now it
sounds hardcoded in the kernel about the internal user, reserved word or
not, its just a bad idea.
joe
Barry Deevey wrote:
As of yet I'm unsure how the application would be affected if I rename the
role - I need to do some investigation.
I tried this in Oracle 8 and it worked fine - It just
Barry,
What you need to do is stop using the INTERNAL role.
Create your own role. Grant access to the tables to this role. And then
grant this role to your application user. Everything should be fine.
As I said, you made a mistake back when you started using the INTERNAL role.
Now that
Yep.
If you remove IPC, then you just can't connect to your database through
sqlnet over IPC.
direct IPC - bequeath will still work.
Tanel.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Tuesday, November 18, 2003 7:24 AM
What will happens if
INTERNAL_USE is an keyword (to enable an undocumented feature) in ALTER
DATABASE
command. THis can be used to convert the database character set if the
existing
char set (national charset) is the superset of the db charset. You can just
run
the ALTER Database command to convert the db charset.
Do you mean with starting the database? Not at all. The db will register
when the listener comes up.
But it might take time to register. Was the interval 3 minutes?
You can use alter system register to force a database to register itself
with listeners configured in init.ora (local_
Yong didn't do it, he merely posted a reply.
Mercadante, Thomas F [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
11/18/2003 06:54 AM
Please respond to ORACLE-L
To:Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Subject:RE: 'internal' role and 9i
Title: col_usage$ question
Does anyone know what this table (sys.col_usage$) is used for? To me it sounds like something that CBO might appreciate ... but any ideas? It is referenced by dbms_stats and dbms_stats_internal packages ...
Thanks in advance
Raj
Here's a cut and paste from the Software Investment Guide.
Notice that Std edition may not be used on a machine *capable*
of utilizing 4 CPU's. It does not matter how many are installed
according to the terms of the license.
The 'Processor Metric' section shows an exception to that which
Hey all,
Just dumped my primary Windohs desktop at work in favor of Gentoo (yay!),
and am installing O9iR2. As I'm browsing thru the Install docs as I am want
to do prior to installs, I see that they recommend using:
echo somevalue /proc/sys/kernel/someparm
...in the startup to set
[EMAIL PROTECTED] uname -a
dgux orion R4.20MU06 generic AViiON PentiumPro
[EMAIL PROTECTED] make love
Make: Don't know how to make love. Stop.
Raj --
That's fbulous!
Thanks,
Bambi.
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Bellow, Bambi
INET: [EMAIL
Yeah,
I realized that afterward - sorry Yong.
Tom Mercadante Oracle Certified Professional
-Original Message-From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]Sent: Tuesday, November 18, 2003
11:24 AMTo: Multiple recipients of list ORACLE-LSubject:
RE: 'internal' role and
That is a new table in Oracle 9 and is used by Oracle to track what columns
are used in predicates. At present the only use of that information that I
am aware of is in the procedure dbms_stats.gather_table_stats ( ...,
method_opt = 'for columns ... size auto');
At 09:29 AM 11/18/2003, you
In 9i on Linux training materials they do recommend sysctl.conf for
permanent settings... Maybe the reason is, that proc file system exists on
all (properly configured) modern linuxes, but sysctl.conf is distribution
specific...
Tanel.
- Original Message -
To: Multiple recipients of
Gopal,
Are you saying with an undocumented parameter or command, I can alter database
(national) character set us7ascii even if my current (national) character set
is utf8?
Yong Huang
--- K Gopalakrishnan [EMAIL PROTECTED] wrote:
INTERNAL_USE is an keyword (to enable an undocumented feature)
Yong,
You said It's not always easy to have a futuristic view to avoid potential
problems. The
developer probably shouldn't have granted anything to internal back then.
But
it wasn't obvious at that time that doing so could cause a problem later.
I totally disagree with you. Your quote implies
How to recycle sbtio.log . Is it same way as
alert.log ?
-ak
On Tue, 18 Nov 2003, Jesse, Rich wrote:
Hey all,
Just dumped my primary Windohs desktop at work in favor of Gentoo (yay!),
and am installing O9iR2. As I'm browsing thru the Install docs as I am want
to do prior to installs, I see that they recommend using:
echo somevalue
Barry,
I suggest you open a Tar with Oracle, unless you're sure changing your
application is easy. Oracle obviously missed this little detail by
over-rejecting a previously legitimate role. If 9i's Release note doesn't say
how to deal with this case, then Oracle support should open a bug.
Tom,
There is no fun ... my favorite quotes used to be in Apple MPW compiler ...
the one I like most is ...
You can't modify a constant, float upstream, win an argument with the IRS,
or satisfy this compiler
Raj
Raj --
I fear this will get off-topic'd pretty quickly, and, I further fear that
you
Raj:
Wolfgang is right. It is populated by SMON (I think every 15mins SMON
flushes the data to COL_USAGE) and the predicate columns are updated
(or collected) from the hard parse of the SQLs.
Best Regards,
K Gopalakrishnan
-Original Message-
Wolfgang Breitling
Sent: Tuesday, November
AK,
Yes. You can simply rename it or delete it. I rename
itmonthly. And then delete it the next month. It truely is a
worthless file.
Tom Mercadante Oracle Certified Professional
-Original Message-From: AK
[mailto:[EMAIL PROTECTED]Sent: Tuesday, November 18, 2003 12:04
INSERT STANDARD DISCLAIMERS
Yes. You can use INTERNAL_USE keyword to convert the database character set.
I think there is a utility called 'csscan' character set scanner which can
be
used to determine the possibility of the INTERNAL_USE conversion.
END DISCLAIMERS
Best Regards,
K
See below:
On 11/18/2003 11:44:25 AM, Jesse, Rich wrote:
Hey all,
Just dumped my primary Windohs desktop at work in favor of Gentoo (yay!),
and am installing O9iR2. As I'm browsing thru the Install docs as I am want
to do prior to installs, I see that they recommend using:
echo
Thanks KG,
Raj
Rajendra dot Jamadagni at nospamespn dot com
All Views expressed in this email are strictly personal.
QOTD: Any clod can have facts, having an opinion is an art !
-Original Message-
Sent:
On 11/18/2003 11:54:41 AM, Bellow, Bambi wrote:
[EMAIL PROTECTED] uname -a
dgux orion R4.20MU06 generic AViiON PentiumPro
[EMAIL PROTECTED] make love
Make: Don't know how to make love. Stop.
Raj --
That's fbulous!
Thanks,
Bambi.
--
Actually, it really is. It shows us
Oracle 9.2.0.4 64-bit
Solaris 9
List:
I've used dbca to create scripts for a new database
using the option Generate Database Creation Scripts.
This completed without errors.
When I vi the init.ora file created by this process,
vi issues a message
init.ora [Incomplete last line] 98 lines
it is a worthless file and what i have done is ln
/dev/null to sbtio.log on my unix box.
this removes the necessity to recycle.
thanks
sai
--- AK [EMAIL PROTECTED] wrote:
How to recycle sbtio.log . Is it same way as
alert.log ?
-ak
--
Please see the official ORACLE-L FAQ:
Gopal,
In case I didn't make my message clear, I wanted to know if using that keyword
allows us to change character set from a superset to a subset (e.g. from UTF8
to US7ASCII). The documented command ALTER DATABASE CHARACTER SET new charset
is only for changing from a subset to a superset. If
Hi,
I had sent this some time back but got no answer for version 8.1.7.For table I understand auditing is an option. What about for index? Thank YouA Joshi [EMAIL PROTECTED] wrote:
Hi,
Is there an easy way to find out if a table or anindex is being used. I mean short of going thru all code or
On Tue, 18 Nov 2003, Mladen Gogala wrote:
And more below:
See below:
On 11/18/2003 11:44:25 AM, Jesse, Rich wrote:
Hey all,
Just dumped my primary Windohs desktop at work in favor of Gentoo (yay!),
and am installing O9iR2. As I'm browsing thru the Install docs as I am want
to do
I think it is complaining about lack of a carriage return at the end of the last line.
Raj
Rajendra dot Jamadagni at nospamespn dot com
All Views expressed in this email are strictly personal.
QOTD: Any clod can have
Thanks Wolfgang.
Raj
Rajendra dot Jamadagni at nospamespn dot com
All Views expressed in this email are strictly personal.
QOTD: Any clod can have facts, having an opinion is an art !
-Original Message-
Jacques,
I checked your example, I think there are some issues here:
1. Original queries provided below do use merge join.
2. We could have missing indexes which can exist on real system.
3. Timings below is not a criteria -- after gathering statistics and
creation an index on val this both
Title: col_usage$ question
Hi!
If you describe this table then you see that this
table stores column usage information in filter and join predicates for database
objects. From describe, you see there are several filter and join conditions
tracked for an object's (obj#) columns (intcol#).
It does. If you do a full import from a US7ASCII database into a
WE8ISO8859P1 database and you need to change the WE8ISO8859P1 database to
US7ASCII then this is the way to do it.
ALTER DATABASE CHARACTER SET internal_use us7ascii;
Make sure you know what you're doing. See Metalink Doc Id
what is max filesize for utl_limit.fopen ? I am
unable to past beyond 51K .
-ak
I have successfully openedmulti megabytefiles ... text files.
Are you storing it somewhere? What does your error stack look
like?
Raj
Rajendra dot Jamadagni at nospamespn dot
com All Views expressed in this
You mean line feed.
Jamadagni, Rajendra [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
11/18/2003 09:59 AM
Please respond to ORACLE-L
To:Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Subject:RE: dbca init.ora incomplete last line
I think it is
AK,
Do you get INVALID_MAXLINESIZE exception? Documentation says it maxes at 32767.
I didn't find file size limit. UTL_FILE.FOPEN also has a max_linesize that can
be set to 32767.
Yong Huang
--- AK [EMAIL PROTECTED] wrote:
what is max filesize for utl_limit.fopen ? I am unable to past beyond
How about \n at the end?
On 11/18/2003 01:54:29 PM, [EMAIL PROTECTED] wrote:
You mean line feed.
Jamadagni, Rajendra [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
11/18/2003 09:59 AM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L [EMAIL
Since making love is simpler, should they start with outsourcing it?
-:)
Igor Neyman, OCP DBA
[EMAIL PROTECTED]
-Original Message-
Mladen Gogala
Sent: Tuesday, November 18, 2003 12:44 PM
To: Multiple recipients of list ORACLE-L
On 11/18/2003 11:54:41 AM, Bellow, Bambi wrote:
[EMAIL
A - The only suggestion I've heard is to take the contents of V$SQL, perform
EXPLAIN PLAN on all SQL, and try to build a list of indexes that are used.
Hardly foolproof. I think Burleson has some scripts in his book Oracle
High-Performance Tuning With STATSPACK, IIRC.
Dennis Williams
DBA
Are you looking to see if statements are using indexes or how often index
blocks are being read?
Daniel Fink
A Joshi wrote:
Hi, I had sent this some time back but got
no answer for version 8.1.7. For table I understand auditing is an option.
What about for index? Thank You
A Joshi [EMAIL
I've never used that, but this syntax seems to work:
alter database character set internal_use us7ascii;
Of course, you could lose some characters or mess up your data completely
that way...
Tanel.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent:
Could the column info be used to 'recommend' indexing?
Daniel Fink
Tanel Poder wrote:
Hi!If
you describe this table then you see that this table stores column usage
information in filter and join predicates for database objects. From describe,
you see there are several filter and join
Well, 'ALTER INDEX MONITORING USAGE' should do the trick. Results should be
in V$OBJECT_USAGE. In connection to that, here is a sweet little bug in oracle 9.2.0.4:
SQL select name from v$fixed_table where name='V$OBJECT_USAGE';
no rows selected
Don't tell that to oracle, they might even fix
Absolutely
At 01:39 PM 11/18/2003, you wrote:
Could the column info be used to 'recommend' indexing?
Daniel Fink
Tanel Poder wrote:
Hi! If you describe this table then you see that this table stores column
usage information in filter and join predicates for database objects.
From describe,
Looking to see if any statement has accessed the index in say 30 days. So basically : "how often index blocks are being read". So I can decide to drop unused indexes. T
Thanks Daniel for your help.
Daniel Fink [EMAIL PROTECTED] wrote:
Are you looking to see if statements are using indexes or how
At 11:04 AM 11/18/2003, you wrote:
Hi!
If you describe this table then you see that this table stores column
usage information in filter and join predicates for database objects. From
describe, you see there are several filter and join conditions tracked for
an object's (obj#) columns
... but the database is 8.1.7 - no monitoring allowed
Mladen Gogala [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
11/18/2003 12:44 PM
Please respond to ORACLE-L
To:Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Subject:Re: ** find whether table or
How about a little upgrade before the end of the day?
On 11/18/2003 03:59:25 PM, [EMAIL PROTECTED] wrote:
... but the database is 8.1.7 - no monitoring allowed
Mladen Gogala [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
11/18/2003 12:44 PM
Please respond to ORACLE-L
Well, it's not a perfect solution but the following query will show you
what is in your buffer pool (v$bh) and match it to each objects total
segment size and give you a report of what is really taking up the buffer
pool and was percent of each object is in memory. It's not a perfect (old
old
This is just an idea, so please test it thoroughly (and then test it again!)
Any and all comments (including "Are you brain-dead, Dan?") are welcome.
How about periodically sampling v$bh for index segment headers? This
assumes that any index access reads the header (true/false?) for the
Daniel,
it
will work but indexes are present in buffer cache also because of updates. The
only possibility is to store each index in questionin separate tablespace
and monitor the i/o. If number of reads will be equal or little bit
greaterthan thenumber of writes than this is a candidate.
Well, 'ALTER INDEX MONITORING USAGE' should do the trick. Results should
be
in V$OBJECT_USAGE. In connection to that, here is a sweet little bug in
oracle 9.2.0.4:
SQL select name from v$fixed_table where name='V$OBJECT_USAGE';
no rows selected
Don't tell that to oracle, they might even
Hi all,
Our app is written using pl/sql's packages. When I
tried to display one form and it runs forever. So I
closed the form, and changed the sql in the package
that generates the form, and recompiled the pkg, and
the recompiling hangs. I checked the database, there
is no locking there. What
I think it means the last line in your file doesn't end in a newline
character.
It shouldn't hurt anything but you should be able to get rid of the
message by adding the newline character.
--Walt Weaver
Bozeman, Montana
On Tue, 2003-11-18 at 09:54, Barbara Baker wrote:
Oracle 9.2.0.4 64-bit
I was too lazy to look for it on asktom.oracle.com, but here's what I read at the site
a while ago (if you search on index usage or something like that you should find Mr.
Kyte's answer). Tom Kyte has the following suggestions:
a) In Oracle 8.0 and earlier - put an index all by itself in a
In constructing my last answer, I looked at v$bh and the class#
column (thinking that perhaps indexes or index headers had their
own class). I came across several different values for the
class# from 0 to 36 (not inclusive). Oracle's doc does not clear
it up and Steve Adam's book's information
Joze,
Excellent Point! I knew there was something wrong about this, I just
could not figure out what it was...
Daniel Fink
Joze Senegacnik wrote:
Daniel,it
will work but indexes are present in buffer cache also because of updates.
The only possibility is to store each index in question in
This was discussed on 31-October and look in the archives for the details, but:
What did you check for locking - if dba_locks or standard utllockt.sql then it won't
show.
You could use Steve Adam's script Executing_packages.sql at
http://www.ixora.com.au/scripts/misc.htm to see what packages
No, differently from table access, index access
doesn't require reading of index segment header. The query reads the root block
directly. Root block address is probably gotten from IND$ table columns file#
and block#+1. The block# represents index segment header location in a datafile,
root
You know, I remember reading the stored outlines trick, but I completely forgot about it.
Nice catch Jacques.
Jacques Kilchoer [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
11/18/2003 02:19 PM
Please respond to ORACLE-L
To:Multiple recipients of list ORACLE-L [EMAIL
I
do.
Raj
-Original Message-From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]Sent: Tuesday, November 18, 2003
1:54 PMTo: Multiple recipients of list ORACLE-LSubject:
RE: dbca init.ora incomplete last lineYou mean line
I think there's another complication in using buffer cache (x$bh or v$bh). If
the index is scanned in parallel, the blocks are not cached in buffer cache.
To avoid counting buffers created due to index update, maybe we can simply say
where v$bh.status in ('CR','READ').
Yong Huang
--- Tanel
Here's the presskit, which gives a decent hi level overview.
http://www.postgresql.org/presskit/en/presskit74.html
One
new feature that is not listed in the 7.4 release is Point In Time Recovery,
which will hopefully make it into the 7.5 release. According to the
PostgreSQL lists, it's in the works.
Best regards,
David B. Wagoner Database Administrator Arsenal Digital
Solutions Web:
help!!!
Implementation: web environment
i'm using orarpp utility to allow user to print rrpt files locally. When the
user choose a printer namely 'prtsvr' that is connected to printer server to
print the rrpt file, orarrpt creates a file at the client side that
initiated the printing instead of
--Boundary-00=_4PUK6RO0
Content-Type: Multipart/Alternative;
boundary=Boundary-00=_4PUK12S0
--Boundary-00=_4PUK12S0
Content-Type: Text/Plain;
charset=shift_jis
Content-Transfer-Encoding: quoted-printable
Joshi,=0D
=0D
Hello!
What do you think of additional backup method as multimaster replication?
Isn't it a way to continue working normal, if main database crash and I move
all connections to another master site?
--
Oracle 9i DBA beginner
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Well, to be helped you should probably tell us something about OS
(Windows, RSX, CP/M, MPE, OS/2, GCOS, DOS/VSE or Ultrix?)
What type of printer connection? (LPD, CUPS, SystemV lp, Amtrak)
What is rrpt? Is it a part of Oracle*Reports or RPG/400? Son of rpt/rpf?
What is the oracle version? What
Hi all,
Looking for possible causes for a ELFCLASS64 version of
/oracle/product/8.1.7/lib/libobk.so accidentally cropping up for a 32-bit install.
This was on a client site ... so some details I can't disclose ... but it's a clean
8.1.7.0.0 install under Solaris (8 I think). Caused the
os: win2k
network : tcp
oracle db version: 817
ias version:9ias rel1
no error messages pops.
The Forms application calls a Report and specifies that it should be output
to file with the
applicable Orarrp extension.
2. The Report runs and places the output on the Web server where it will be
ldd?
On 2003.11.19 00:19, Grant Allen wrote:
Hi all,
Looking for possible causes for a ELFCLASS64 version of
/oracle/product/8.1.7/lib/libobk.so accidentally cropping up for a 32-bit
install. This was on a client site ... so some details I can't disclose ...
but it's a clean 8.1.7.0.0 install
We are doing Interest processing on a SET of 10,000 Bank A/cs
SQL Query :-
select field names,rowid into :b0,:b1,...
from TBA_ENTITY_INTEREST_TBL
where (entity_id=:b105 and entity_type=:b106)
for update of same (above) field names
nowait;
Above SQL Query Shows DIFFERING values (taken
VirVit,
You haven't specified Oracle version and your tolerance for data loss. If
you can afford to lose the data in the most current redo log file, then I
would suggest 9i Data Guard in maximum performance mode (or Standby in 8i)
solution over the MM replication. The DG solution does not affect
90 matches
Mail list logo