Title: RE: Oh Dear, Mr. Ellison's at it again
19th
Century... electric cars were being made in the late 1800s... not much progress
since.
:
)
Pat.
-Original Message-From: Goulet, Dick
[mailto:[EMAIL PROTECTED]Sent: Wednesday, April 02, 2003 4:04
PMTo: Multiple recipients
Ellison: Linux Will Wipe Microsoft Out Of The Data Center
In dramatic terms, Oracle Corp. Chairman and CEO Larry Ellison
predicted the open-source operating system will wipe Microsoft off the
face of the earth in the battle for the data center market.
Title: RE: Oh Dear, Mr. Ellison's at it again
-Original Message-
From: Goulet, Dick [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, April 02, 2003 10:45 AM
To: Multiple recipients of list ORACLE-L
Subject: Oh Dear, Mr. Ellison's at it again
Ellison: Linux Will Wipe Microsoft
Title: RE: Oh Dear, Mr. Ellison's at it again
Jeremy,
True, you should watch what you wish for because it may
well come true!!
RIP Oracle???
Dick
Goulet
-Original Message-From: Jeremy Pulcifer
[mailto:[EMAIL PROTECTED]Sent: Wednesday, April 02, 2003
2:14 PMTo: Multiple
of list ORACLE-L
Subject: Oh Dear, Mr. Ellison's at it again
Wonder if anyone has clued him into MySql and PostGreSql yet!!
Dick Goulet
Senior Oracle DBA
Oracle Certified 8i DBA
--
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Weaver, Walt
INET: [EMAIL
You mean that MySQL and Postgres will wipe out Microsoft? :)
-Original Message-
Sent: Wednesday, April 02, 2003 1:45 PM
To: Multiple recipients of list ORACLE-L
Ellison: Linux Will Wipe Microsoft Out Of The Data Center
In dramatic terms, Oracle Corp. Chairman and CEO Larry Ellison
, April 02, 2003 11:45 AM
To: Multiple recipients of list ORACLE-L
Subject: Oh Dear, Mr. Ellison's at it again
Wonder if anyone has clued him into MySql and PostGreSql yet!!
Dick Goulet
Senior Oracle DBA
Oracle Certified 8i DBA
--
--
Please see the official ORACLE-L FAQ: http
Just had a thought here, have not tried it yet. I have a database that I am
working with that is generating 28 GB of redo each day. I would really like
to know what objects are generating all this redo without going through the
hassle of mining a bunch of log files. It occurred to me that if
Ethan - 28-gig GAAAK!
Ideas:
1. Sample your SQL buffer to start getting some ideas.
2. Use LogMiner to read some of the archive logs to see the DML
statements. You can also directly see how much redo is being generated by
each statement.
I think you have the right idea, probably some
Hi,
The simpler approach is to check the user level redo (or session level
redo) using the v$sysstat,sesstat views and you can find the programmes
associated with those huge (!) redo.
Dumping the redologs and analyzing is just complex when you have a
simple solution ;)
=
Have a nice day
It might work to turn on monitoring on the tables.
alter table xyz monitoring;
Then periodically check dba_tab_modifications.
-Original Message-
The goal is to identify the objects, then identify the jobs
that work on
those objects and see if I can reduce redo. I suspect a lot
Which version of Oracle ?
Has someone switched on supplemental logging
at the database level, perhaps ?
Have you got dbms_job kicking in every 5 seconds
with job_queues set to 10 ? (Honest, I have seen
it happen, and the effect on redo was astonishing -
and there was only one job actually ever
Ethan,
monitoring is active and my stats are up to date I should be able to
multiply the total number of updates, inserts and commits by
the average row
size and get a rough % of what objects are generating the most redo.
Note that the amount of redo does not depend on the average row
So here are two takes at the problem, one takes a look at costly (in regards
to amount of redo) tables and the other indexes. Note this is only a way to
guestimate this information.
select owner,
table_name,
round((ratio_to_report(ttl) over ()) * 100, 1) as percent_ratio
from (
Title: RE: Oh Where Oh Where Is My Redo Coming From
Ethan,
I think focusing on which transaction generates more redo will be more helpful than which object ... right?
Let me know if I didn't understand your question completely ...
Raj
Indexes on such tables which has DML...
Regards
Rafiq
Reply-To: [EMAIL PROTECTED]
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Date: Fri, 21 Feb 2003 07:44:18 -0800
Just had a thought here, have not tried it yet. I have a database that I am
working with that is generating
Ethan,
v$sess_io will provide a list of the sessions generating block
changes and, therefore, redo. Link this back to v$session, etc. for the
'offending' sessions, sql. I've used this several times with great
success. Sometimes it was a data load, others it was bad sql.
When you change a
Yes, that is what I was saying, however large rows or tables with a lot of
indexes would also be prone to generate more redo, that is why I suggest
joining DBA_TAB_MODIFICATIONS to DBA_TABLES to get avg_row_len and
DBA_IND_COLUMNS to get the total # of columns indexes on the table, the
thought
Sure that is the way I would typically do it, but in this case I have an
application that is running 8000 batch processes per day, redo is very
consistent for most of the 24 hours. I asked myself what is the simplest
way to figure out which objects likely generate all of this redo.
Monitoring
Title: RE: Oh Where Oh Where Is My Redo Coming From
Since SharePlex for Oracle and LiveReorg are dependant on redo log volumes, transaction sizes, and things like that we have developed a free utility that parses through the redo logs for some pre-defined amount of time, and let's you know
John that doesn't make sense to me. Deleting 10,000 rows from a table with
200 columns is certainly going to generate more redo than a table with 1
column (which I am sure you know, so there must be some confusion in my
understanding). The avg_row_len is going to be much bigger in the table
with
So what about an app server process that has been handling requests for 2
weeks? In this case I still won't know anything about the process(es)
causing the redo.
I have monitoring scripts that can trigger execution of another script(s) to
enact SQL trace on the top N sessions generating redo,
Title: RE: Oh Where Oh Where Is My Redo Coming From
Ethan,
There is a option in TOAD that automatically commits after every query ... turn that off ...
Raj
-
Rajendra dot Jamadagni at espn dot com
Any views expressed here
in the line of over 60 per minute (log in, do something,
log out,
oh and commit between each row of course). That on top of hundreds
of users
accessing the system and oh 4 or 5 developers working in the same
database
doing whatever they please. I found one guy the other day had issued
290,000 commits
Naw, A beltway bandit who has a contract at WRight-Pittiful with Aeronautical
Systems.
Dick Goulet
Reply Separator
Author: Charlie Mengler [EMAIL PROTECTED]
Date: 6/26/2002 8:23 AM
I suspect that this position will be at Wright-Patterson Air Force
Apparently another
plane has crashed in New York, a 767 has gone down in Queens according to news
reports over here.
The information contained in this communication is
confidential, is intended only for the use of the recipient
named above, and may be legally privileged. If the reader
of
It's an American Airlines A-300 Airbus, leaving JFK, bound for Santo
Domingo, Dominican Republic.
Robertson Lee - lerobe wrote:
Apparently another plane has crashed in New York, a 767 has gone down
in Queens according to news reports over here.
--
Please see the official ORACLE-L FAQ:
It was
an Airbus A300 and apparently suffered engine failure.
-Original Message-From: Robertson Lee - lerobe
Sent: 12 November 2001 14:41To: ORACLE LIST
(E-mail)Subject: Oh no
Apparently another
plane has crashed in New York, a 767 has gone down in Queens according
Para: Multiple recipients of list ORACLE-L
Asunto: Re: Oh no
It's an American Airlines A-300 Airbus, leaving JFK, bound for Santo
Domingo, Dominican Republic.
Robertson Lee - lerobe wrote:
Apparently another plane has crashed in New York, a 767 has gone down
in Queens according to news
29 matches
Mail list logo