/nlzanen1/External/MEY/NL)
Sent by: Subject: RE: Free Buffer Waits
[EMAIL PROTECTED
the batch job started because that starts right
after startup. So basically the system (with no other logins as per
listener.ora file). had been waiting about 6-7 hours since startup on free
buffer waits.
We did not introduce any other jobs to the batch and it has been running
normally since that one
Hi All,
As mysteriously as it took over 6 hours to run yesterday, it was back at 35
minutes today. Free buffer waits dropped out of the top 10 again as well.
My question now is:
If nothing changed to the database (we did not modify anything because we
could not find the source of all evil
Pretty interesting. Maybe it's possible, but a simpler explanation is
that something about your workload either during or maybe immediately
before your batch job was different.
The most common case in which I've seen 'free buffer waits' waits is
when all the forces of nature combine to make
Hi All,
We have a production database that has a batch job running on it for months
now. Last night one part of the batch job that normally takes between 20-30
minutes took well over 6 hours.
The only difference I can see between today other days is that the Free
Buffer Wait event was the top
It's in centiseconds (100th sec) in all versions
up to 9i. In 9i, some have changed to (1th sec) IIRC.
So, if you're not in 9i, each of your waits is one second
(100.8971 * 0.01) and it waited a total of around 22000 secs
(between 7 and 8 hours). Bad.
As for the reasons, they can be many
free buffer waits waits indicate that your DBWR can't keep up with its
workload. Often caused by inefficient SQL competing with DBWR for an I/O
device.
Cary Millsap
Hotsos Enterprises, Ltd.
[EMAIL PROTECTED]
http://www.hotsos.com
-Original Message-
Zanen
Sent: Thursday, June 13, 2002
Sounds like the write latency on your storage is high or you have an abusive. 'free buffer waits' is the db writer failing to flush it's cache fast enough to disk.
George
On Thursday, March 7, 2002, at 06:39 PM, Manytrees wrote:
Hello all,
Does anybody know what parameters I should
waits
Hello all,
Does anybody know what parameters I should be
tuning/change to try reduce the number of busy buffer waits. I
have a system which at times has over 15+session waiting on "free buffer
waits".
The explanation that I have been able to find
Hello all,
Does anybody know what parameters I should be
tuning/change to try reduce the number of busy buffer waits. I have
a system which at times has over 15+session waiting on "free buffer
waits".
The explanation that I have been able to find so
far is that session a
Hello all,
Does anybody know what parameters I should be
tuning/change to try reduce the number of busy buffer waits. I have
a system which at times has over 15+session waiting on "free buffer
waits".
The explanation that I have been able to find so
far is that session a
:[EMAIL PROTECTED]]
Reply To: [EMAIL PROTECTED]
Sent: Monday, December 17, 2001 4:30 PM
To: Multiple recipients of list ORACLE-L
Subject: RE: free buffer waits DISK I/O
Can't answer your question but sounds like you might benefit from reading
Gaja Krishna Vaidyanatha's
HI list.
I was tuning a database (Oracle 7.3.4) that was
waiting for free buffer waits mostly.
I tuned DBWn batch size properly (increased it till
4MB) as well as buffer cache size (it's 600MB now)
too and I got some improvement but I didn't get a
major improvement.
But some weeks ago, when we
Pablo,
Free buffer waits is normally due to DBWR not being able to write out
modified blocks to disk fast enough.
By changing the disk layout you have eliminated disk contention and hot
spots and have enabled DBWR to write out the dirty buffers to disk faster
thus making enough free buffers
event 10046 and also by looking at
session_event
both show free buffer waits and the table is being read SEQUENTIALLY many
times
!! (FTS should be done using SCATTERED READS )
even when doing scattered reads, the blocks read are 4 or 5 at a time...(not
16)
other issues
- i did a count
Rahul,
Again Free Buffer wait does not directly
correspond to the database buffer cache.
This can happen even if you have umpteen
free buffers in the buffer cache.
This generally happens during FTS of a
large table , which is what happening in
your case. If you have primary key defined
for
list, (AIX, 7.3.2)
i have found another abnomality which my db is suddenly showing..
when doing a count(*) on a million row table the session is
waiting on free buffer waits !! i have 80M db buffers and most of them are
empty/free !!
all the free buffer waits are on one of the db file
17 matches
Mail list logo
