Message -
From:
Hemant
K Chitale
To: Multiple recipients of list ORACLE-L
Sent: Monday, December 01, 2003 5:14
PM
Subject: Re: SESSION_CACHED_CURSORS --
RE: Parse Vs Execute
CURSOR_SPACE_FOR_TIME is FALSE.This is an
Oracle Apps R11 install.HemantAt 05:29 AM 30-11-03
Jonathan,
I've understood that when cursor_space_for_time is true, then unpin is only
done when cursor is closed, thus there's no need for pinning/unpinning for
every execution of a cursor. This should reduce hits on library cache
latches since pinning is not done so often?
Hermant,
I've
To: Multiple recipients of list ORACLE-L
Sent: Monday, December 01, 2003 5:14 PM
Subject: Re: SESSION_CACHED_CURSORS -- RE: Parse Vs Execute
CURSOR_SPACE_FOR_TIME is FALSE.
This is an Oracle Apps R11 install.
Hemant
At 05:29 AM 30-11-03 -0800, you wrote:
What's the value for your
-
From: Hemant K Chitale
To: Multiple recipients of
list ORACLE-L
Sent: Monday, December 01, 2003 5:14 PM
Subject: Re: SESSION_CACHED_CURSORS -- RE: Parse Vs Execute
CURSOR_SPACE_FOR_TIME is FALSE.
This is an Oracle Apps R11 install.
Hemant
At 05:29 AM
That was my understanding, too. The problem with unpinning only at
the specific close is that smon cannot free shared pool memory belonging
to the cursor if the cursor is pinned, so the shared pool usage skyrockets.
The only way to circumvent the problem is to set CURSOR_SHARING to FORCE.
That
Mladen,
I don't think it's SMON who is coalescing free memory extents. I'm not
entirely sure here, but I think if any server process explicitly frees a
freeable chunk, then the 16-byte header of immediate next chunk is checked,
if this is also free both chunks are coalesced and header of next
: SESSION_CACHED_CURSORS -- RE: Parse Vs Execute
Mladen,
I don't think it's SMON who is coalescing free memory extents. I'm not
entirely sure here, but I think if any server process explicitly frees a
freeable chunk, then the 16-byte header of immediate next chunk is checked
Steve Adams talks about AST's which are blast from the VMS past.
More exactly, AST used to stand for Asynchronous System Trap,
and was processed in a similar way like signals on the unix.
The CPU IPL was elevated to AST delivery level (IPL 2) and
AST was queued on the target process entry. When
It's not being the case. I would really, really like to
know how does Oracle implement AST's?
There's no such thing you won't find from Ixora:
http://www.ixora.com.au/q+a/misc.htm
Search for AST :)
Tanel.
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Tanel
I was the guy who asked that question long time ago, but I'm not sure
how exactly are sockets used. Socket is, essentially, a pipe. You must
have someone reading and someone writing it. That is not exactly what I'd
call an AST.
On 12/02/2003 01:39:28 PM, Tanel Poder wrote:
It's not being the
I have to admit that I wasn't thinking about replying
to your comment when I sent this email. However,
I think you are correct - there is an effect of extra
items not being releasable from the shared pool
when cursor_space_for_time is true. (From memory
of one of Steve's seminars, it is the Heap
CURSOR_SPACE_FOR_TIME is FALSE.
This is an Oracle Apps R11 install.
Hemant
At 05:29 AM 30-11-03 -0800, you wrote:
What's
the value for your cursor_space_for_time parameter?
Tanel.
- Original Message -
From: Hemant K
Chitale
To: Multiple recipients of
list ORACLE-L
Sent: Sunday,
What's the value for your cursor_space_for_time
parameter?
Tanel.
- Original Message -
From:
Hemant
K Chitale
To: Multiple recipients of list ORACLE-L
Sent: Sunday, November 30, 2003 8:54
AM
Subject: SESSION_CACHED_CURSORS -- RE:
Parse Vs Execute
I
You still have to hit the library cache to execute
a statement as it needs to be pinned in share mode,
and unpinned when you finish with it. Library cache
latch waits can be a symptom of excessive executions.
Have you checked the library cache latch children
to see if the load is evenly
14 matches
Mail list logo