Hi,
Check whether SMON is active ? What is your tablespace status ? Fragmented ? If yes, SMON would be trying to coalesce. Check for the PCTINCREASE at the tablespace level ... if it is greater than 0.. you will have to let SMON clear the fragmentation at tablespace level.
EVEN if you shutdown and
Greetings, insomniacs and readers in other time zones.
For those of you trying to get bugs resolved, a new patch set might
contain your fix.
8.1.7.2.1 for NT is out - 100MB - and here's my rebuild script so far
...
Paul
(Yes, I *have* to apply every patch set as soon as it comes out)
importa
I think Steve is probably a little more precise
in his statement. But there are definitely cases
where using before-row triggers is more expensive
that using after-row triggers to do the same job.
The cost appears as extra current mode gets
and redo log generation.
Thomas Kyte has a demonstratio
That's true provided you're not using any of the new
features introduced in 8.0, and later versions, that
automatically cause the CBO to be used (e.g.
partitioning, degree or instances > 1, IOT, etc...)
When in doubt, I always check metalink note: 66484.1.
HTH,
-- Anita
--- Jon Walthour <[EMAI
Jay,
The size of the buffer cache and the number of
DB_WRITER_PROCESSES has a big impact on the time it
takes to truncate a table.
Each DBWR process has to scan its portion of the
buffer cache to flush any blocks belonging to the
truncated table or its indexes. This is a sequential
process as o
Chris,
I'm confused by your statement "there is no
materialized view functionality for SE" as this is
completely incorrect.
Read-only snapshots (called materialized views in 8i)
have ALWAYS been available with Standard Edition
(well, at least since 7.3) as they're considered Basic
Replication.
IMHO,Problem in this situation is that ,in 7.3.4 ,if I am not wrong ,thier
optimizer must be RULE as default and after migration if optimizer is not
explicitly set as RULE , default optimizer in 8.1.6 is CHOOSE. This is the
reason their objects never analyzed in 7.3.4 or now in 8.1.6.
If optimi
That's an interesting thought, but why do you
think that only one DBWR process can scan
at a time ?
If you have multiple working data sets,
surely you can have one dbwr process
per working data set doing the scanning.
After all, each working data set has its
own 'range reuse' list (or should that
Mohammad,
For this particular case you're probably correct.
What I was attempting to explain, and not doing very
well, was that the absence of statistics does NOT mean
the RBO will ALWAYS be used. Many people don't
realize this and assume that because they have
OPTIMIZER_MODE = RULE that they w
Kumanan,
Yes, you can easily do this as long as you have the
datafiles for the system tablespace and the datafiles
for any RBS tablespaces used by online RBS's.
This procedure is documented in metalink note: 96197.1
"Recovering a Dropped Table from a Full Database
Backup"
The recovery process w
Jonathan,
I believe it's to prevent other DBWR processes from
attempting to modify blocks belonging to the truncated
object until all of them have had a chance to flush
the blocks, but I honestly don't remember all the
details. I'll dig up my notes when I'm back in the
office and post them to th
Eric:
Thanks for the attachment.
Do you have 3 of 4 and 2 of 4?
-Original Message-
Sent: Friday, August 17, 2001 7:37 PM
To: Multiple recipients of list ORACLE-L
4 of 4
10K
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Norwood Bradly A
INET: [EMAIL
Anita,
You are 100% right for your explanation. I observed that behaviour just last
week when we migrated our second production database from 7.3.4.5 to
8.1.6.2 and my colleague forgot to define optimizer as rule in our new
init.ora file and resultantly optimizer became CHOOSE and from very ne
Alex:
This is a fallacy that's been around since Oracle 6. In actuality, one
wants to set the initial and next extents of a temporary tablespace to
the same size as the size of the average sort in memory to ensure that
most disk sorts can be accommodated within a single temporary tablespace
exten
Anita,
As you might have guessed from my poast earlier on the ORACLE-L, I
applied the 8.1.7.2.1 patch set to the 81715 EE server last night (this
morning).
It had no effect - still receieved the same ORA-00600.
I even set compatible=8.1.7.2.0 and bounced - no effect.
I was working from home yest
To find the maximum number of concurrent sorts from instance start, run
the following query:
SELECT LEAST(disk_sorts, peak) AS peak_concurrent
FROM (SELECT VALUE AS disk_sorts
FROM sys.v_$sysstat
WHERE name = 'sorts (disk)')
, (SELECT /*+ ordered */
SUM(l.
On Friday 17 August 2001 07:36, Kevin Lange wrote:
> I have been collecting commodores and Apples for a while to keep in my own
> PC museum. I have been looking for a good working TRS-80.
Too bad I divested myself of most of my Radio Shack computers.
Model 16 running Xenix
Model 3 with 48k
Jay,
Index blocks can be reused, but unlike a table block,
they are not reused until *all* data is removed from
the block.
So if you remove most rows, it won't help if most blocks
still have a pointer to a row.
Jared
On Wednesday 15 August 2001 13:48, Miller, Jay wrote:
> I'm puzzled.
> We h
Having used RedHat, Slackware, Mandrake, SuSE and others, my
vote goes for SuSE.
Excellent distribution, good installation and admin utilities, certified
for Oracle also.
Jared
On Friday 17 August 2001 05:23, Rao, UmaSankara S (CAP) wrote:
> Hi Gurus,
>
> Can u please tell me the most stable v
let's see your script.
Jared
On Wednesday 15 August 2001 17:35, David Turner wrote:
> I was wondering how most people handle linefeeds in columns. My first
> shot at a perl dump script left one of my files with multiple lines per
> record because there we linefeeds in one of the columns.
>
> Ho
http://www.crtnet.com/servlet/page?_pageid=262,264&_dad=prod&_schema=PORTAL30&_mode=3
This makes me want to take a whack at Portal.
Paul
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Paul Drake
INET: [EMAIL PROTECTED]
Fat City Network Services-- (858) 538-505
What is interesting and not surprising is that for all the sites listed with the
longest average
running uptime, not one of them use Microsoft.
Ron Thomas
Hypercom, Inc
[EMAIL PROTECTED]
"Either lead by example, or become a terrible warning"
I have a 12.5 million row table for which a truncate has been running for
>2hrs. What can cause this?
(Solaris 2.6, Oracle 8.1.7)
Thanks in advance,
- Paul
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Troiano, Paul (CAP, GEFA)
INET: [EMAIL PROTECTED]
Fat City
23 matches
Mail list logo
