Before starting svrmgr set the oracle_sid to test or prod
In command window:
SET ORACLE_SID=TEST
HTH
Volker Schoen
E-Mail: mailto:[EMAIL PROTECTED]
http://www.inplan.de
-Ursprüngliche Nachricht-
Von: Naveen Nahata [mailto:[EMAIL PROTECTED]]
Gesendet: Montag, 2. September 2002 19:08
Hi Volker,
That is OK. I can even start with STARTUP PFILE = 'appropriate_pfile';
But in this case I am unable to start it using NT Services. That is the
problem.
How do I get both the DBs to start by starting the services.
I don't know why suddenly the problem has started coming. All
System Admin, thought you'd like at least one more :)
joe
RO-Hosur wrote:
Please don't sent any more mail to this ID
Please take it as a request
Please ignore my name
Mail ID : [EMAIL PROTECTED]
Thanks
System Admin
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Have you tried to recreate your TEST service using :-
Oradim -delete
-sid TEST
Oradim -new
-sid TEST
-intpwd password
-startmode auto -pfile=d:\directory\testini.ora
Lewis Bishop
---
Barclays Enable/ISS/OPTS - Oracle OCP Database Consultant
Phone - 020 8298 3418
Did you check the registry settings for HKLM\Software\Oracle and
HKLM\System\currentcontrolset\services?
Volker Schoen
E-Mail: mailto:[EMAIL PROTECTED]
http://www.inplan.de
-Ursprüngliche Nachricht-
Von: Naveen Nahata [mailto:[EMAIL PROTECTED]]
Gesendet: Dienstag, 3. September 2002
Well, the flood of responses (not) to this topic probably answers one of the
points raised!
While endorsing all that Dennis has stated, I would just like to add
something.
Most crucially, replication is an exercise in logic, which fundamentally
depends on getting your database design correct on
Hello
all,
I
get a process id by calling getProcessId()..I want to know from program whether
that processId exists.
Any way to find out
how ?
Thanks and
regards,
Santosh
HI ALL,
I am using RMAN with the catalog on a NT Box and 2 targets databases on
Netware boxes.
First a made a whole backup on one database (it was OK), after I changed
the target database on the command line of RMAN.
I got this error :
RMAN-03002: failure during compilation of
Lewis,
Ya that is the only logical thing left to do. Haven't done it as yet because
all the registry parameters seemed fine and ORADIM does nothing more than
creating a few registry entries.
But ORADIM knows more than me what registry parameters to set, so i'll give
it a shot.
Regards,
Naveen
-Please don't sent any more mail to this ID
-Please take it as a request
-Please ignore my name
There is no excape. Leaving the list is futile. Your efforts will be crushed. You
will be assimilated into the Borg. You shall be called one of none.
The Borg
-Original Message-
Sent:
Forgot to mention in the previous post.
Rather DELETE and then CREATE, I think the best will be to edit the INSTANCE
using ORADIM, isn't it?
Naveen
-Original Message-
Sent: Tuesday, September 03, 2002 4:33 PM
To: Multiple recipients of list ORACLE-L
Have you tried to recreate your
Hello
Processes can test the validity of PID's by signalling the PID
with a ZERO signal number. To signal the process you use the
kill() system call. When a zero signal number is used error checking
is done but no signal is sent so that you don't have to worry
about what the process is going to
in the shell enter echo $$. didn't indicate in what application.
kr mr
[EMAIL PROTECTED] 09/03/02 13:29 PM
Hello all,
I get a process id by calling getProcessId()..I want to know from
program whether that processId exists.
Any way to find out how ?
Thanks and regards,
Santosh
--
Please
Thanks a lot to all who have contributed their experiences and ideas for
thsi problem.
I have to look into the application and business details whether this is
feasible or not. Have to talk to our tech head.
Marul.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL
Personally, I never bother with edit... oradim is pretty simplistic...
that's just me though...
Lewis Bishop
---
Barclays Enable/ISS/OPTS - Oracle OCP Database Consultant
Phone - 020 8298 3418
Mobile - 07950 380857
Email - [EMAIL PROTECTED]
-Original Message-
Sent: 03 September 2002
Naveen,
I would try edit first. If it doesn't work, then delete and create. It is
no big deal - delete does not touch the database files at all - just the
windows directory.
Tom Mercadante
Oracle Certified Professional
-Original Message-
Sent: Tuesday, September 03, 2002 8:28 AM
To:
Been a while since I used them, but I had to go through the file and remove
some non-printable characters. You can do this pretty quickly with any
number of text utilities.
Do yourself a favor and use RMAN for backups. You can script any variety of
backups fairly easily once you work your
If adding 400 records slows down the system, I guess it is time to take a
hard look at your set-up and see where the problems are occurring by
measuring and analyzing the waits in the database.
How did you establish that balancing indexes causes the delay? No matter
what kind of solution you
Concerning gqlplus
I spoke to the author about this issue and he is supposed to be addressing
it. I still find it useful and hope he fixes this problem soon.
-D-
On Fri, 30 Aug 2002 12:58:26 -0800
[EMAIL PROTECTED] wrote:
Just tried it.
Took 5 minutes to break it with this:
set
-- Santosh Varma [EMAIL PROTECTED]
Hello all,
I get a process id by calling getProcessId()..I want to know from
program whether that processId exists.
Any way to find out how ?
Depends on the O/s. Simplest way is to grep the output of
ps. On systems with a /proc file system (e.g.,
Now that it is public,
Mazel Tov!
-Ari
-Original Message-
Carmichael
Sent: Saturday, August 31, 2002 6:53 AM
To: Multiple recipients of list ORACLE-L
sigh, this is what comes of answering mail when I am tired... this was
meant to go private.
Not that I didn't want to share this
Ya, Thanx everyone for the ideas.
I think the problem must be the creation of password files, because after
that you have to register the password file with the instance using ORADIM.
The problem was not big but what and why is more important to me now
Thanx
Naveen
-Original
Naveen,
Ther password file has nothing to do with ORADIM. ORADIM is used to create,
modify or delete NT services for maintaining the Oracle instance on the NT
box.
I'm not sure what you mean by register the password file...using ORADIM.
ORADIM can be used to create the service using a
hello
hey - this is a great and really comprehensive explanation.
[EMAIL PROTECTED] 09/03/02 14:36 PM
Hello
Processes can test the validity of PID's by signalling the PID
with a ZERO signal number. To signal the process you use the
kill() system call. When a zero signal number is used
time for me to ask the experts again.
My data warehouse will be 9.2, with all locally managed tablespaces. We
will be following what I have taken to calling the Goldilocks
principle -- that of small, medium and large tablespace extent sizes,
with variations in that we will separate indexes and
Rachel,
Good news should not be kept quite. And kids are a pile of fun, especially
when they aren't yours. 8-)
Dick (GrandPa) Goulet
Reply Separator
Author: Rachel Carmichael [EMAIL PROTECTED]
Date: 8/31/2002 3:53 AM
sigh, this is what comes
Rachel - I have been using LMTs with uniform extents (Oracle-recommended
variation) for a couple of years now with Oracle 8.1.6, and now 9.2. The
Oracle-recommended extent sizes are 128k, 4m, and 128m. As far as PCTFREE,
et. al., these are at the table level, so my opinion would be that the
HI all,
Do you've any pl/sql for detect redundant
index?
thanks,
Adriano Freire
I stumbled on this new book the other day and bought it. I think it is
terrific. We DBAs are often asked to help tune SQL statements. This book
offers lots of ideas for fixing SQL statements. Most thorough explanation of
how the rule-based and cost-based optimizers work that I have encountered.
We used this method (I wish I had come up with the Goldilocks code name...)
in 8.0.4 on a couple of data warehouses. The bottom line is that we had 1
space related failure (application code filled up the error log) in over 9
months of operation. Which so impressed management that they decided to
Dennis, who creates tables in your DB? If devs do, don't you worry that one
could accidentally create a 512MB table in your 128K TS, instead of a 512KB
one?
I really would like to implement LMTs here, and am doing so in certain
restricted instances where I'm the only one who creates the objects
Hi,
We are having an Oracle 8.1.7 database with Dataguard 2.6 on Sun Solaris.
Just wanted to check, if I am required to add a datafile to the Standby Datbase
Manually, if I add a datafile to ROLLBACK Segment Tablespace in the Primary database.
Thanks for your time and help.
CP
Rachel,
You did not say if you would be using the UNIFORM option for the
LMT's. If you allow the system to choose the initial sizing there can be
a lot of wasted space as the table size grows. The system will choose
sizing options that you most likely would not choose. I can't find my
reference
If you add a file to a primary, you must add it to the standby, reguardless
of what object had the file.
-Original Message-
Sent: Tuesday, September 03, 2002 1:04 PM
To: Multiple recipients of list ORACLE-L
Hi,
We are having an Oracle 8.1.7 database with Dataguard 2.6 on Sun Solaris.
We may end up reworking the extent sizes, right now they are smaller
than those but we are still in stage one, haven't gone live yet (and I
don't even want to think about what a pain it will be to change things
when we do go live).
Data load test coming up soon, so I'll have a better idea of
Mark Gurry. that kinda says it all. :)
--- DENNIS WILLIAMS [EMAIL PROTECTED] wrote:
I stumbled on this new book the other day and bought it. I think it
is
terrific. We DBAs are often asked to help tune SQL statements. This
book
offers lots of ideas for fixing SQL statements. Most thorough
Thanks Grant
I have already done this,
and this is what I got:
SET oracle-l MAIL
You are already receiving all mailing list messages. In order to hold the
messages temporarily, issue a SET listname NOMAIL command to ListGuru.
SET oracle-l REPRO
You are already copied on all messages which
Dan,
Why 3 extent sizes in each tablespace? I can see how you prevented the
dreaded fragmentation problem by making them multiples of one another
but I don't understand why you did that.
If I'm using LMTs can I still do that? I haven't really used them, I'd
thought to not specify the initial
Ron,
yes it helps, thanks! I forgot to mention that I will be using UNIFORM
sizes for the LMTs, I don't really see the advantage to autoallocate
I can probably get away with the defaults for the parameters (if I
don't use automatic segment space allocation) but it will waste some
space in some
I recall that these were the days before LMT and at the start of moving away
from fine-grained object sizing. In hindsight, 1 size per tablespace makes
more sense. There could be several exceptions. For example, all of the
reference tables/indexes were in a single tablespace. Since these were
On Tablespace Creation In sys.dba_tablespaces , Field ALLOCATION_TYPE Defaults to
SYSTEM
. Hence NEXT_EXTENT of Created Tables is NOT Taken from the User-specified Value but
internally by Oracle itself . This leads to Excessively Large Number of EXTENTs (Small
in Size)
To Allow Table
Rachel
The varchar business on your fact table worries me. I'm not trying to be
critical, but to bring up some issues you may want to consider before you
begin the big load. Unfortunately too many DWs end up with a flawed data
model (ours included) that limit its usability. Sometimes the
I
presume 'redundant' means that a column in the leading position of an index is
also in the leading position of another index.
Here
is a quick and dirty SQL script to generate this information. Unfortunately, it
repeats the information, but it does give you the
information.
SQLbreak on
So, if this is the goldilocks approach...who are the 3 bears?
duh. small tablespace (Baby Bear), medium tablespace (Mama Bear) and
large tablespace (Papa Bear)
and if a table is Goldilocks then one and only one tablespace will be
just right
Rachel
--- Fink, Dan [EMAIL PROTECTED] wrote:
I
Rich - Good point. Yes, I create all the tables here, at least in
production, and I probably wouldn't use autoextend if the situation were
otherwise. The other thing to consider is if you are using uniform extents,
by definition you have bought into the philosophy that you can have many
extents
Hi Friends,
My all background processors accessing same file system, I did some reorg
this weekend, But I shut down database like 4 times...all processors are
waitingAny help will be highly appreicated!!!
oracle@baan1 $ ps -ef|grep 14816
oracle 14816 1 2 13:59:00 - 0:49
You have to state the directory (full path) in a format statement. Then
rman will put it where you want it.
Ruth
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Friday, August 30, 2002 10:28 PM
Hello,
I took an RMAN backup of a database (for
The question posed was not whether extent management local should be used, but
whether automatic segment space management should be used.
As this is a data warehouse, I would not expect you to have transactions trying to
change the same block. Assuming you are loading; that is, inserting
Hi Friends,
My all background processors accessing same file system, I did some reorg
this weekend, But I shut down database like 4 times...all processors are
waitingAny help will be highly appreicated!!!
oracle@baan1 $ ps -ef|grep 14816
oracle 14816 1 2 13:59:00 - 0:49
my fault, the customer table is a dimension table, not a fact table.
Doh!
--- DENNIS WILLIAMS [EMAIL PROTECTED] wrote:
Rachel
The varchar business on your fact table worries me. I'm not
trying to be
critical, but to bring up some issues you may want to consider before
you
begin the
Can someone tell me where to start to find out how to access an MS Access
table from Oracle. Basically what I want to do is be connected to an Oracle
database in SQL*Plus and execute a query against an MS-Access table. Is
this possible?
Thanks,
Jeff
--
Please see the official ORACLE-L FAQ:
My one of disk showing full active...very hot!!! by bground processes
accessing...
hdisk51 88.7 926.6 139.3 2780 8
hdisk51 85.5 1217.0 176.2 3660 0
hdisk51 91.8 980.2 149.3 2948
From: Peter R [EMAIL PROTECTED]
Came across this very slanted view of Unix/Linux vs
Windows. Hope it comes in handy for anyone doing
research on this particular topic.
http://www.cuug.ab.ca/~leblancj/nt_to_unix.html
mkb
__
Do You Yahoo!?
Yahoo! Finance - Get real-time stock
Search on Metalink for 'Heterogenous Services'. The documents there
describe how to setup a dblink to an MS Access database.
HTH,
Beth
-Original Message-
Sent: Tuesday, September 03, 2002 4:39 PM
To: Multiple recipients of list ORACLE-L
Can someone tell me where to start to find
My database okay, its slow, I can't do much work, too much waitingI want
to bring as normal processing...
From: Fink, Dan [EMAIL PROTECTED]
To: '[EMAIL PROTECTED]' [EMAIL PROTECTED]
Subject: RE: Pls respond...: URGENT
Date: Tue, 3 Sep 2002 13:58:02 -0600
Peter,
What is the problem?
What kind of reorgs did you do?
-Original Message-
Sent: Tuesday, September 03, 2002 4:59 PM
To: Multiple recipients of list ORACLE-L
My one of disk showing full active...very hot!!! by bground processes
accessing...
hdisk51 88.7 926.6 139.3 2780 8
The ArcServe agent has a tendency to login to your database
and never logout, eventually eating up all your processes or
logins if you don't keep an eye on it.
It also tries to backup TEMPORARY tablespaces by putting
them in backup mode, generating an ORA-3217 in the process.
FYI: We're
Thanks Ian, that was indeed the question, although the other
information is useful as well.
if I understand what you are saying correctly, it works but it won't
really buy me anything and I might do better controlling the space
myself. And while you have had no problems, you've heard negative
Is it possible there is a full analyze running on the database? You said
you did
a reorg this weekend. If you bounced it 4 times it could be going through
backout
and recovery. Either one would keep it pretty busy.
R. Smith
-Original Message-
Sent: Tuesday, September 03, 2002 4:08 PM
Check out note # 116565.1 on MetaLink.
It should clear it up.
Jared
On Saturday 31 August 2002 02:28, kommareddy sreenivasa wrote:
Hello all,
DB: 8i
OS: soalris 2.8
why DBMS_SPACE.FREE_BLOCKS is not giving correct
status:
create table ram1 (a number);
analyze table ram1 compute
I do backup database nightly using export utility. When I restore database
using import, it takes more than 6 hours to finish. Is there a way to speed
up import process? Please advise.
Thanks,
David
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Nguyen, David M
Doc ID: 109730.1
-Original Message-
Sent: Tuesday, September 03, 2002 5:08 PM
To: Multiple recipients of list ORACLE-L
Search on Metalink for 'Heterogenous Services'. The documents there
describe how to setup a dblink to an MS Access database.
HTH,
Beth
-Original Message-
There are quite a few restrictions on function-based indexes.
The Oracle SQL guide lists them all. Have you checked to
ensure that you're following all the rules?
Jared
On Saturday 31 August 2002 07:53, Marul Mehta wrote:
Even after giving the hint its not working.
I guess you can't have
I'm managing an OPS
configuration (4x HP 9000/N, HP-UX 11/64, RDBMS
8.1.7.1)
and I'm having an
application dependency on a temporal order of sequence
numbers.
With OPS that
becomes a problem because each node caches a set of sequence
numbers
(20 by default).
Oracle has an option,
Are you importing to existing tables with indexes ?? Are you importing the
indexes ??
One big way to speed up an import is to NOT import any indexes and, if the
objects exist on the target database, to delete any indexes and rebuild them
in a sperate action after the import is complete.
I agree with you Dennis, I've already bought it.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Tuesday, September 03, 2002 12:34 PM
I stumbled on this new book the other day and bought it. I think it is
terrific. We DBAs are often asked to
I agree with you Dennis, I've already bought it.
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Tuesday, September 03, 2002 12:34 PM
I stumbled on this new book the other day and bought it. I think it is
terrific. We DBAs are often asked to
It
looks like when option "ORDER" is used Oracle guarantees the generated values
will be in order since the "CACHE" option will be ignored by Oracle even if it
was requested.
This
is in the parallel mode.
Look
at note: Note:1031850.6
Waleed
-Original Message-From: Gogala,
So, proper LMT means no LBE? ;) Great analogy! All the head-spinning and
the green projectile vomiting and such...
BTW, yes that is a good paper. I've read it and am trying to deal with the
extent sizes as it applies to our DB, as only about two dozen of the 800+
tables are larger than 128MB
I have one huge table (takes up about 30% of the total database storage)
which has a monthly batch deletion of old data. I had PCTFREE and PCTUSED
set to the defaults of 10 and 40 respectively.
I occurred to me that I could probably free up a lot more space by
increasing the PCTUSED so that
1- Delete means freeing space.
2- When the used space in the blocks falls below PCTUSED, the block needs to
go back to the free list.
3- This requires extra processing cost.
4- So lowering PCTUSED will lower the frequency of triggering this
procedure.
Waleed
-Original Message-
Sent:
I see 2 different addreses for the same hash_value in
v$sqlarea.
Why do I see this?
Here's an example:
hash_value address
-- ---
3749804 4064082C
3749804 4192941C
I wonder if someone can shed some light on this
Txs
If you run OPS and specify order, it works like no cache.
My question to you: Why cripple OPS and your business performance by having
this requirement ? Spending a few bucks to get rid of this dependency will
improve the performance, until you run in to the next problem ;-)
Anjo.
On
Yes, but when analyzed, it turns out that NOCACHE will also
yield ordered results. What I'm interested in are internal differences
in behavior. My assumption is that with ORDER oracle queries the instances
directly, while NOCACHE will simply read/write everything from the disk.
On 2002.09.03
Could because of different child cursors ?!
On Wednesday 04 September 2002 01:04, you wrote:
I see 2 different addreses for the same hash_value in
v$sqlarea.
Why do I see this?
Here's an example:
hash_value address
-- ---
3749804 4064082C
3749804
I run a script to grant privileges on database objects to developers and it
normally completes in a few minutes. Lately, the script started hanging and
getting ORA 4021 after 5 minutes. Discovered the problem is a pin lock.
The sessions holding the locks are not even active, they had accessed
The way I see it is: If you specify ORDER then the only way Oracle can
enforce this is getting it from the dictionary which means no caching will
be implemented.
If you need the data to be ordered then (in my opinion) it's better to
declare what you need by using option ORDER.
Using option
Unfortunately, we have an application dependency and I was required
to come up with a quick dirty fix. Thanks for your reply.
On 2002.09.03 19:10 Anjo Kolk wrote:
If you run OPS and specify order, it works like no cache.
My question to you: Why cripple OPS and your business performance
Title: RE: PCTUSED - when is block added to freelist?
The way I understand it:
If you have a low pctused, then you have less blocks being moved to the freelist (because it's less probably that a block will be moved to the freelist): so reduced processing costs during update (if a row length
If I understand Jay's question correctly, what he's asking is
not how PCTUSED and PCTFREE work, but what action or
actions trigger Oracle to put a block back on the freelist after
changing PCTUSED to a higher value?
Is that correct Jay?
Jared
On Tuesday 03 September 2002 15:38, Miller, Jay
Exactly.
Ian MacGregor
Stanford Linear Accelerator Center
[EMAIL PROTECTED]
-Original Message-
Sent: Tuesday, September 03, 2002 2:23 PM
To: Multiple recipients of list ORACLE-L
Thanks Ian, that was indeed the question, although the other
information is useful as well.
if I
While Oracle says you can have unlimited extents, practically it
doesn't hand more than at MOST between 1000 and 4000 extents in a
table.
And you can just adjust your baby bear, mama bear and papa bear extent
sizes in all your tablespaces, no one says the extent sizes are fixed!
Geez, years
I agree with Anoj, you need to talk to the business folks to remove this
dependency. Else you may encounter waits/queues on getting the next sequence
numbers. One of the benfits in OPS and in RAC is the sequence cache option,
because each instance will not have to query the Oracle's fast
Jared,
i hope u hv taken care of veritas nbu licenses. the resellers try to apply 1 Oracle
agent license for each database instance instead of 1 license for each server machine.
-Mandar
-Original Message-
From: Jared Still [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 03,
Try RE-Booting the Machine , if you think the behaviour is very abnormal
If you are on IBM AIX , ensure that you have applied the the 1 off patch for fsync
available on 8.1.7.4 (32 Bit) OR 8.1.7.2 (64 Bit)
Setting Multiple db_writers might help , after turning disk_async_io = false .
O.S.
if exp/imp were used Ensure that Statistics are DELETED ( RE-Created , if needed) as
they will be highly incorrect
Check for DEGREE 1 of the Tables Indexes . This will also Cause Optimizer to
choose BAD Plans if NO Statistics Exist
-Original Message-
Sent: Wednesday, September
list,
i have been learning about OID for the last couple of weeks..
i have managed to install and run the ldap server on my w2k, i can run the
OID manager GUI, and can also do a ldap://wb-rahul:4034; from my IE5
now i want to pupulate the directory with some new info, and test the
storage
One note: 9i automatic segment space management does not automate PCTFREE;
that still functions as before. It does cause PCTUSED, FREELISTS, and
FREELIST GROUPS to be ignored, however...
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Tuesday,
Mladen,
Is there any way to have developers/users access
the sequence via a function, instead of accessing the sequence
directly?
If so, then perhaps you could modify the sequence
to addthe temporal component, while maintaining the use of a cached
sequence for uniqueness? Such as:
SQL
Title: RE: Constraints problem
Hi,
Is it
possible to change the sysdate? (Solaris 5.8, Oracle 8.1.7)
TIA.
K.
Hi All,
We have a table which can contain more than half a
million records. When we try to insert some 10k records in the empty table it
get inserted in 10 min. but as the size increases time taken to insert also
increases. After 350,000 records it takes around an hour to insert 10k
92 matches
Mail list logo