One of the teams here is planning to run anything
from 8 to 16 database instances [no indication on
sizing yet, but gut-feel SGAs are 200MB to 1GB
and DB sizes 500MB to 40GB] on a large server,
something like a Sun E6800 or an equivalent
HP or Fujitsu server.
1. How many of you do run, and are
Hi All,
I've the following qry :
select distinct(decode(b.sys_id,'TRDENG',a.cust_bas_no,a.cosmos_base_no))
baseno ,a.br_cod,pty_nam,bank_name
from piar_fr_psd a, psd b, bank_br c
where a.psd_id=b.psd_id
and a.psd_serial_num = b.psd_serial_no
and b.bank_id = c.bank_id
and (((sys_id =
We have got a requirement to keep all the documents on repository. We are
looking on the possibility of keeping documents on Database or filesysytem.
Could you please let me know your views which one I should go for.
Regards,
Dhanvir
I have run/am running multiple oracle databases on a single server on
AIX/Linux/W2K/NT although not with sga's in the GB area.Just make sure you
get away with your I/O requirements
We only had seperate homes for seperate oracle versions, so all 817
databases were in one home and 8.0.5 databases
Could anybody of the list give me the correct explanation of the below ?mechanism provided by ORACLE for table replication - Snapshots and SNAPSHOT LOGs
Thanks and Regards,
Santosh
Select * from (
select
distinct(decode(b.sys_id,'TRDENG',a.cust_bas_no,a.cosmos_base_no))
baseno ,a.br_cod,pty_nam,bank_name,
rank() over (partition by baseno ,a.br_cod order by pty_nam) as rk
from piar_fr_psd a, psd b, bank_br c
where a.psd_id=b.psd_id
and a.psd_serial_num = b.psd_serial_no
and
Fresh tablespace Created specifically to create a testcase for an iTAR.
-Original Message-
Sent: Monday, March 10, 2003 4:30 PM
To: Multiple recipients of list ORACLE-L
Was this a fresh tablespace or were there swiss cheese holes available to
fill?
Ron Thomas
Hypercom, Inc
[EMAIL
The version is 8.1.7.1.0
The report is via this query:
SELECT partition_name, extent_id, bytes/1024, bytes/1024/1024
FROM dba_extents
WHERE segment_name = 'FORMATTER_DATA_HISTORY'
ANDowner = 'KEVIN'
ORDER BY 1, 2
The file is not autoextent.
The table is populated by
Read this,
http://download-west.oracle.com/docs/cd/A87860_01/doc/server.817/a76965/c31repli.htm#12888
Regards,
Kamaljeet Singh
(NCDB ASG)
MBT, 52
BarrackSquare, Martlesham. IP5 3RF.
Off. 01473 667170
Mob. 077 5368 5370
[EMAIL PROTECTED]
-Original Message-
On Mon, Mar 10, 2003 at 08:28:50PM -0800, Charles Hart wrote:
I installed 9.2 on redhat 8.0 with great luck. I installed on a 700 PC
with 756 megs of memory. I was using forms 6i running on NT client and
could not tell when I was pointed at this instance verus one running on a
HP machine.
Title: RE: Oracle position on hints
Hi,
We recently upgraded from 7.3.4 to 8.7 (management plans on getting to that Y2K problem shortly ;-) We had an SQL statement that really needed a hint in 7.3.4. After upgrading to 8.7, I removed the hint and it runs much faster without it. I spend
I think you're right Tim. It was kinda the point I so clumsily tried to
make. There really is no such thing as a direct connection per se.
Everything connects using SqlNet.
-Original Message-
Sent: Monday, March 10, 2003 6:19 PM
To: Multiple recipients of list ORACLE-L
But it still
All -
I am having a problem with the datafiles in a temporary tablespace. I
need to move and rename three different datafiles in the tablespace. I
am able to take them offline - no problem. I cna make the changes at
the OS level. I am running on Unix. But I can't get the changes to
show up
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of DENNIS
WILLIAMS
Sent: Monday, March 10, 2003 15:45
To: Multiple recipients of list ORACLE-L
Subject: RE: Quick question on cursor resource use
Fuzzy
What is prompting you to increase OPEN_CURSORS?
After spending 2 ½ days trying hints, init parameters, re-writing the query, a
completely useless TAR, etc.
to get a query that runs in 1 second on 8.1.6.X to go faster than 1 minute on
9.2.0.2
I found a new to 9.2.0.X dynamic init parameter optimizer_dynamic_sampling, if I
understand it
Hi,
On one of the 7 development box (aix 4.3.3) we have 27 instances Oracle
8172.
All using the same oracle_home.
I can't say it's the fastest response time ;-)
As for the licensing we have a mix of CPU and user licences.
Stephane Paquette
Administrateur de bases de donnees
Database
I can't remember the last time that I saw a server (other than WinNT)
running a single instance. My experience is that it's quite normal to run
multiple instances on a single server.
One Oracle home per version of Oracle. I'm not sure what the point of a
separate Oracle home per instance would
Title: RE: Running multiple instances on a [large] server
On our development RAC servers we have 24 instances ... on each side, performance is okay.
Raj
-
Rajendra dot Jamadagni at espn dot com
Any views expressed here are strictly
I need to determine whether or not a couple of indexes need to be rebuilt.
The problem is the indexes are quite large and on a 24x7 high volume
database. If I try to run an analyze validate structure to gather the data
I need to make that decision, it sets a lock on the table for about an hour
Tom:
You are absolutely right. What I do is to have a nightly script to check the
available disk space (using df -k and awk) and total sum of dba_temp_files
to make sure there always enough disk space for the LMT temp tablespace.
Guang
-Original Message-
From: [EMAIL PROTECTED]
Title: RE: LMT monitoring
Connor,
What on earth you are doing on this list immediately after your Wedding? Which cruise liner has internet access?? I think Disney has ...
ps: Thanks for the algorithm, let me implement and see how good my data dictionary holds up.
Raj
LeRoy,
I just struggled with this last week. You can't move Temporary Data files.
You need to drop and recreate the TEMP tablespace - creating the data files
in the correct directories.
On a side note - here is an interesting feature. When Oracle creates files
for the TEMP tablespace, it does
Just installed Perl 5.8. Ran ppm3 to install the DBD package
andPPM couldn'tfind it. I looked at activeperl site, it stated
dbd for Oracle failed. I have now installed perl 5.6 and can find dbd/dbi
files.
With Perl 5.8, are you supposed to use something other than dbd/dbi to
access an
We are running 11 instances on WinNT with no problems! At one point, we had
26 instances. they are all development instances, so volumn and load are
low. but they all share one Oracle Home.
Tom Mercadante
Oracle Certified Professional
-Original Message-
Sent: Tuesday, March 11, 2003
Not that this helps Raj much, but the algorithm does
vary if the initial size of the segment is large,
along the lines of:
case
when initial_extent 1m then
case when extents 16 then next = 64k,
when extents 80 then next = 1m,
when extents 200 then next
Rathi
Here are some considerations:
1. How many documents / how much storage?
2. Do you need to manipulate the documents while they are stored?
3. Consider Oracle ifs http://www.orafaq.org/faqifs.htm
4. What are your performance requirements?
Dennis Williams
DBA, 40%OCP, 100% DBA
We also run several instances on just about every server we have. Most, but not
all share the same Oracle_home. I only create new ones for newer versions of
Oracle, but for each database I prefer to use seperate mount points such as:
/test/system
/test/rbs
/test/temp
/prod1/system
/prod1/rbs
Title: RE: Running multiple instances on a [large] server
I see some values 0 for chaint_cnt in
dba_tables . How do I know if this is chained rows or migrated rows ?
Any hits .
Thanks,
ak
Thomas,
Now that's an interesting comment. I can't remember ever seeing more than
one instance on an NT machine either except when we were playing around boy
did that crash in a big hurry (NT 4.0 Oracle 8.0). On the other hand I've set
up Linux with 4 instances (same piece of hardware) and
You can do
1. Create a new temporary tablespace (say temp1)
2. Assign all user's TEMPORARY_TABLESPACE to temp1
3. Drop the old temporary tablespace, remove it's OS files.
4. Re-create the temporary tablespace using new OS file names
5. Re-Assign all user's TEMPORARY_TABLESPACE to this new
If you have
millions of documents it may be better to have them stored in the db. If
you have terabytes of documents that are pretty much static/read-only then you
may want to have a separate db for the images and one for your
data. [EMAIL PROTECTED] 03/11/03 04:58AM
We have got a
Hemant
What you are considering is certainly feasible. Consider how compatible
these applications are, whether they have similar requirements in terms of
uptime. Are their performance requirements compatible? One factor to
consider is future upgrade paths of the applications. We seem to run
I'm trying to build a case for management that we need additional DBAs so
I'd like to take a quick poll if I may. What is the ratio of Oracle
databases to DBAs in your shop? This includes development and production
databases. At our shop it's 33:1.
--
Please see the official ORACLE-L FAQ:
I tried it both ways and got the same result.
According to my notes from the Oracle Backup and Recovery course
when you apply your redo logs, the control file gets rolled forward
along with the data files. I believe that the only time you would get
into trouble here is if you had made structural
Title: RE: Analyzing indexes
Chuck,
Do you think these indexes are corrupt? Validate structure doesn't give you statistics like Compute Statistics or Estimate Statistics does.
Jerry Whittle
ASIFICS DBA
NCI Information Systems Inc.
[EMAIL PROTECTED]
618-622-4145
-Original
Hemant
Dick makes a good point about physical memory. The problem with a lot of
instances is that they all have fixed amounts of memory (except in 9i, up to
SGA_MAX_SIZE), and if one instance needs more memory you can't reallocate
unless you bounce it to add memory and maybe bounce several
We've run into this with Backtrack 3.3, which is also supposed to support
8.1.7.x. If I recall correctly, the temp tablespace is there, but offline.
We usually have the create scripts for the database in question, so we just
drop the tablespace and recreate it. We usually create it as locally
Hi All,
I have a after database logon trigger on server/instance 1. All it does is
insert a record into instance on server2 via a database link.
This creates a session on server2. So every user who logs on to server1 a
session is created on server2. The session does not
go away. Does anyone
At RMOUG last week, I heard that someone (Chip Briggs I think) had 5
(count 'em 5) isntances of Oracle running on his PC. I don't think it
was Linux and it definitely wasn't production anything
Rachel
--- [EMAIL PROTECTED] wrote:
Thomas,
Now that's an interesting comment. I can't
ak
Smart-alek answer: Apply one of the methods to eliminate migrated rows,
and if the problem doesn't go away, you know you have some chained rows ;-)
Chained rows are a little difficult to diagnose. Look at the value for
avg_row_len - is it near the db_block_size? I haven't tried this, but
Chuck,
At this location we have a total of 5 databases and 1 DBA. Time off
without an electronic teather is a thing of the past. My last location
was 4 production databases with 1 DBA and a backup DBA from the
development arena of 2 development databases and 1 DBA.
Ron
[EMAIL PROTECTED]
AK,
hi, row chaining is because the big row(row length is bigger than the
block size), while row migration is because of small pctfree and updated rows.
So, just look at the length of the rows. If chained rows, no
way(unless you use larger block size and the row
Classic case of doing a proper requirements analysis. It will cover issues
already raised, but will place those suggestions into the correct context of
your business objectives. There is no cook-book answer - the answer will be
unique to your installation and your environment. If it happens to be
3:1
Production, Development, my personal test DB : Me
Craig Healey
-Original Message-
From: Chuck Hamilton [mailto:[EMAIL PROTECTED]
Sent: 11 March 2003 14:59
To: Multiple recipients of list ORACLE-L
Subject: POLL: Database to DBA ratio
I'm trying to build a case for
Title: RE: LMT monitoring
Here is my interpretation of algorithm suggested by Conner,
(I'll get to others too)
/*
CASE WHEN initial_extent 1m THEN
CASE WHEN EXTENTS 16 THEN NEXT = 64k,
WHEN EXTENTS 80 THEN NEXT = 1m,
WHEN EXTENTS 200 THEN NEXT = 8m,
ELSE NEXT = 64m
WHEN
You can use: execute immediate 'alter session close database link
utilities_itport02_dblink';
After you're done with the insertion
Regards,
Waleed
-Original Message-
Sent: Tuesday, March 11, 2003 10:54 AM
To: Multiple recipients of list ORACLE-L
Hi All,
I have a after database logon
Title: RE: need help with dynamic sql
if you at-least know the source and destination tables will have same number of columns and data types ... you could just do something like
insert into target.table
select from source.table
Raj
Not with execute immediate. Use dbms_sql
I have done this.
For example and help
http://asktom.oracle.com/pls/ask/f?p=4950:8:104566430235951855::NO::F495
0_P8_DISPLAYID,F4950_P8_CRITERIA:584023239495,
Regards,
Kamaljeet Singh (NCDB ASG)
MBT, 52 Barrack Square, Martlesham. IP5 3RF.
Off. 01473
I do have multiple databases on a Win2K server. No problems except that
MicroSoft just loves to write the SGA out to the swap file (about 630M of
SGA, 1.3G of RAM and disk i/o out the wazzoo). I don't know if there was
some specific reason that we never ran more than one instance on WinNT 4.0
33 DBA's to one database? That's pretty danged good! :)
Seriously, this is a question that comes up every so often within Oracle
as well, as clients ask Oracle employees what should be expected.
There's no simple answer to it. As you might expect, the answer is it
depends. :)
Depends on (a
We have 10 development databases that I support directly plus 5
implementation/testing databases that I support as second level and one
production snapshot datamart.
16:1
However, there are also DBAs who support each project. That would make the
ratio closer to 4:1 or 3:1.
Hi,
We're having
slow response on some batch jobs in our production system. A straight SQL
trace shows intermittant slow response times on selects against a variety of
tables. What is odd is:
the statistics are
current for the tables, frequently the results of a
compute.
The same
I wouldn't hesitate to set it to 500 or a even 1000 if your appliation really
needs this many. But if you think about 10's of thousands, I would start
really looking af the application to understand if it were written correctly.
The major memory is not allocated by setting the parameter, but
I will have to check your question on 'fewer sorts'.
As far as changes to the SQL statements, there were 1 or 2 sql statements
changed in the application because they no longer worked efficiently with
the change to 8.1.7. But, on the whole (lets say 98%+ of the time) no
changes were made.
I need to truncate and import data into several schemas. The tables have
lots of constraints. I can produce a script to disable and enable the
contriants but I would like to know more about the constraint_type field in
dba_constraints and what are all the SYS_ contraints? Should I disable all
Title: RE: LMT monitoring
can't create a database with oracle 8.1.7.the installation went
fine, but when i start dbassist on the 2% of thedatabase creation it
tels me: "not connected to oracle". any ideas?
Title: RE: Running multiple instances on a [large] server
We run upto 22 instances on large SUN boxes... Each environment has their own oracle home, own file system, and in some cases-their own UNIX account for the environment..
One loses a little space for the multiple oracle homes, and it
27 databases
to 1 dba, mixed dev, test, and prod. [EMAIL PROTECTED]
03/11/03 11:29AM 3:1 Production, Development, my personal
test DB : MeCraig Healey -Original Message-
href="">mailto:[EMAIL PROTECTED] Sent:
11 March 2003 14:59 To: Multiple recipients of list ORACLE-L
build a case
Title: RE: AFTER database logon trigger keeps sessions open
Try putting a
dbms_session.close_database_link('utilities_itport02_dblink');
after the insert. also aren't we missing a commit?? I'd also make this a autonomous transaction ...
Raj
Dan;
Everything remained the same on the DB except the version. No changes
made.
-Original Message-
Sent: Monday, March 10, 2003 5:34 PM
To: Multiple recipients of list ORACLE-L
Kevin,
The ORA-01555 errors are almost certainly symptoms, not the actual
problem. It has been a few
i have precompiled the cernel with ntfs write
support.
now how can i mount my ntfs partitions to try
it?
RS On Mon, Mar 10, 2003 at 08:28:50PM -0800, Charles Hart wrote:
I installed 9.2 on redhat 8.0 with great luck. I installed on a 700 PC
with 756 megs of memory. I was using forms 6i running on NT client and
could not tell when I was pointed at this instance verus one running on a
HP
How about trying these:
SVRMGR set autorecovery on
Autorecovery ON
SVRMGR recover database using backup controlfile;
Ping
Northrop Grumman IT
-Original Message-
Sent: Tuesday, March 11, 2003 10:34 AM
To: Multiple recipients of list ORACLE-L
I tried it both ways and got the same
Chuck
I think this is a slippery argument. It may appeal to managers but can
lead to bad decisions long-term. Does anyone remember years ago when
managers would measure COBOL programmers LOC (lines of code)? Just measure
how many lines of code each programmer writes each day and vola! you know
[EMAIL PROTECTED] 03/11/03 09:59AM
I'm trying to build a case for management that we need additional DBAs
so
I'd like to take a quick poll if I may. What is the ratio of Oracle
databases to DBAs in your shop? This includes development and
production
databases. At our shop it's 33:1.
It
Seems like we are starting Yet Another Bind Discussion!
Anyway, the peeking into bind variables is basically done when a hard parse
is done. Hence, if multiple sessions execute the same shared SQL statement,
only the first one will actually do the peek and hence, the optimization
for all
6:2
six DBAs to one development + one testing VLDB. Good luck.
-Original Message-
Sent: Tuesday, March 11, 2003 11:29 AM
To: Multiple recipients of list ORACLE-L
3:1
Production, Development, my personal test DB : Me
Craig Healey
-Original Message-
From: Chuck Hamilton
Tom Kyte has an excellent method of finding which are chained and which are migrated
published in a recent issue of Oracle magazine. I don't know which one, possibly the
latest, but will check when I get back in the office. In the meantime, the
information may be on his web site.
A good
volume of database structural changes (new tables, changes to existing
ones) is another to add to that list
--- Pete Sharman [EMAIL PROTECTED] wrote:
33 DBA's to one database? That's pretty danged good! :)
Seriously, this is a question that comes up every so often within
Oracle
as well,
Title: RE: LMT monitoring
I had
the same problem,You have to generate the script with dbassist then run the
script.
-Original Message-From: Milen Pankov
[mailto:[EMAIL PROTECTED]Sent: Tuesday, March 11, 2003 9:00
AMTo: Multiple recipients of list ORACLE-LSubject: can't
Title: RE: LMT monitoring
Is
this on Linux?
If
yes, then it sounds like you need to install the glibc stubs
patch.
If no,
then I don't know what the problem might be.
-Original Message-From: Milen Pankov
[mailto:[EMAIL PROTECTED]Sent: Tuesday, March 11, 2003 10:00
AMTo:
Title: Can't run full exports due to SMON wait
Running 8.1.7.4 on RS6000/AiX 4.3.3. We noticed this weekend that all our
full exports performed via sqlbacktrack for a given database were hanging.
Killing the exports and restarting does not work.
Checking waits, it looks as though the full
SYS_ contraints are the ones that when you created them you did not give a
name. Oracle will automatically assign a name like 'SYS_1234567'.
You can drop tables in the schema before import. Import will automatically
create tables if it does not exist.
HTH.
Guang
-Original Message-
Chuck,
We've got near as many instances as you (40 at last count) but in three
different locations(2 in Mass although there 10 miles apart and one in CA).
Consequently there are three DBA's. You also might want to take a head count of
what I refer to as screaming mimesis (end users who
Title: RE: Oracle Developer Kit 9iV2 on Solaris - Refined
Actually what are best practices for where you put this stuff - apparently some of it can be either on 9ias or on the RDBMS. If you have an application that will be using both - which is best practice? Are there issues related to
I doubt there are "some incorrectly applications", I think that is a rule
rather than an exception.
Raj
-
Rajendra dot Jamadagni at espn dot
com Any views expressed here are
strictly personal. QOTD: Any clod can
have facts,
Title: RE: Oracle Developer Kit 9iV2 on Solaris
As a standard we install web servers separate from RDBMS servers (i.e. on separate hosts).
9iV2 comes with Development Kit.
Is any of that needed on the RDBMS server?
Would I want any of that installed on the web server host on Solaris if my
I currently have 5 development databases and 1 beta production database
that I administer, so that's 6:1, but I'm also the Sys Admin for these
servers as well. These are also very small databases.
robin
-Original Message-
Sent: Tuesday, March 11, 2003 11:44 AM
To: Multiple recipients
Ron,
SYS_ constraints are NOT NULL constraints.
They are also those constraints placed on tables by you that you have not
named. This is why we should always name our constraints!
The Constraint_Type field is decoded as follows:
C = Check Constraint
P = Primary Key
R = Relational (Foreign Key)
28 Oracle (dev, test, qa prod)
3 SQL Server (arent I lucky?
gag)
31:1
Todd Carlson
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Gene
Sais
Sent: Tuesday, March 11, 2003
11:00 AM
To: Multiple recipients of list
ORACLE-L
Subject: RE:
On Tue, Mar 11, 2003 at 09:00:09AM -0800, Milen Pankov wrote:
RE: LMT monitoringcan't create a database with oracle 8.1.7.
the installation went fine, but when i start dbassist on the 2% of the
database creation it tels me:
not connected to oracle.
any ideas?
--
look in the alert log
Last place I was at was 35:1 but we had top notch guys and
and organized environment..The kinds of databases you have..
what kinds of work being done on them etc.
the high availability and service level agreement and change
control requirments.
Now we have mega-sensitive instances.. needing Now
On the Oracle side there's 2 of us (myself and one other) supporting 20
databases, so that makes us 10:1.
However, in addition to this there is also the IBM mainframe/Software AG
Adabas side where it is my privilege to support the 5 legacy database
environments by myself.
Jim Damiano
--
Please
Title: Oracle Database audit...
This may or may not be of value...
Under the Security:
Possibly users with the oracle account password on the machine?
This would obviously allow people to conn / as sysdba.
Maybe? :)
Have a good one!
RE: Analyzing indexesNo but validate structure populates the index_stats
view which is the only way I know of to get the index height, leaf rows,
deleted leaf rows, and pct. of used space which is what I normally use in
determining if an index needs to be rebuilt or not. Is there another way?
We currently have twenty-one databases, thirteen production and eight which are
either test or development. Of the thirteen production databases five need to be up
on a 24 X 7 basis; physics experiments and or accelerator monitoring depend on them.
Nine of the production databases are
Do all share the same address space or does each instance get it's own copy
of Oracle.exe and it's own 2g address space?
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Tuesday, March 11, 2003 10:04 AM
We are running 11 instances on WinNT with no
thats the problem, I do not know for sure that the tables are identical in each
schema. Sorry, I left that out. I have to check for that before I even get to that
point. If there are 'NOT NULL' columns in the destination table, then I throw insert
to an error table, if not then I build the
Title: Oracle Database audit...
Oops.
This wasn't meant to go to the list.
sorry
-Original Message-
From: Chris Stephens
[mailto:[EMAIL PROTECTED]
Sent: Tuesday, March 11, 2003
12:20 PM
To: Multiple recipients of list
ORACLE-L
Subject: Oracle Database audit...
This
RE: Running multiple instances on a [large] serverBy environment do you
mean instance? If so, couldn't you accomplish the same thing with one
oracle home on each server for each version installed? That's what we do in
our clustered environments. Each node has an oracle home for each version of
Consider yourself lucky. We've got a 71:1 ratio here.
Chuck Hamilton
I had the same error message on NT with Oracle 9.2 recently.
The first problem was that the init_sid.ora file was created a directory different
from where the db create scripts were looking for it. The second problem was that one
of the init parameters was for Enterprise Edition and I was
A simple num of databases to DBA ratio is quite meaningless.
You need to take into consideration of at least the following:
o Size of the database, transaction volume. What kind of up time your
database
needs to provide.
o Complexity of the database and application which run on top of it.
I have written two programs to deal with this:
1) a plsql procedure that shows all the LIO for a chained row system wide.
This works like sar:
ser serveroutput on
execute oraperf.analyze_lio(10, 1);
2) an object row chainer analyzer, this will find which SQL statements and
Chuck,
I support 17 development, 2 production oracle databases, 3 DB2/MVS databases
and 1 SQL Server. A coworker supports only 1 application and that is in
UAT/Dev environment but it is a highly active application in the
multi-Terabyte range (UDB). We have external prod. support for some of
139 remote production servers/instances (30 are large to VLDB, the rest
small), 5 development instances, 1 9iAS test instance, 3 dba's. Mostly
reporting with batch ETL jobs run nightly.
Scott Shafer
San Antonio, TX
210.581.6217
-Original Message-
From: Robin Ilardi [SMTP:[EMAIL
31:1 will be 40:1 by the end of the year. I have a second guy in training as the
second DBA.
3:1 oracle app servers
23 24X7 medical databases of one type or annother
8 Dev/Test databases
1 production app server
2 dev/test app servers
...JIM...
[EMAIL PROTECTED] 3/11/03 8:59:20 AM
I'm
Get DBI and DBD::Oracle from www.xmlproj.com/PPM, courtesy of Ilya Sterin.
Jared
Jeffrey Beckstrom [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/11/2003 06:34 AM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
cc:
Chuck,
That's not really a good measure. It really depends on
a number of things.
* number of users
* number of apps
* size of db's
* quality of apps ( some require more work than others )
* amount of new development
* lots more I can't think of at the moment...
Best thing to do is document
Chuck,
I've been convinced that rebuilding indexes is a waste of time.
In fact, it can cost you time, as rebuilding indexes can kill your
peformance while the indexes again seek their 'level'.
Check into at asktom.oracle.com. There's some good examples.
jared
Chuck Hamilton [EMAIL
1 - 100 of 157 matches
Mail list logo