Notice that the show sga command output doesn't contain a line saying "shared pool size", but instead says "variable size". That's because it's the size of the variable part of the SGA, which includes the row/dictionary cache, the library cache (usually together referred to as the shared pool) plus a number of other structures including the hash buckets and the buffer header structures of the buffer cache, the lru/lruw/whatever structures and more. Therefor Variable Size will always be bigger than shared pool size.
I hope this answered somebody's question somewhere ;-)
Mogens
Jonathan Lewis wrote:
I hadn't heard the historic explanation before,
so I'll pass on that.
As far as the 16MB is concerned - I believe
the free memory includes any free space
left in the shared_pool_reserved_size.
Since the shared_pool_reserved_size defaults
to 5% of the shared_pool_size (I think) it isn't
necessarily a surprise that you have 16MB
of free memory when your shared_pool size if
320MB. (On the other hand, is the reserved
size supposed to be extracted from the main
pool, or additional too the main pool)
The latching thing is always good for a cop-out.
I suspect that v$sgastat would become a major
bottle neck if it were always latched and updated
in real time. So it seems very likely that it would
always be wrong.
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
Coming soon a new one-day tutorial:
Cost Based Optimisation
(see http://www.jlcomp.demon.co.uk/tutorial.html )
Next Seminar dates:
(see http://www.jlcomp.demon.co.uk/seminar.html )
____England______January 21/23
The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html
-----Original Message-----
To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
Date: 02 January 2003 15:13
Sorry for being so vague, but sometimes I can't help it...
It was my understanding in the Oracle7 days that the name of the
statistic "free memory" was actually a verb and a noun (i.e. as in
"free Nelson Mandela" or "free Willy"), and the number shown alongside
this statistic was the cumulative number of bytes freed in the Shared
Pool. In other words, every time "N" bytes were freed from the Shared
Pool, then the statistic was incremented by "N". At least, this
explanation would have accounted for the absurdly huge numbers seen in
the V$SGASTAT view for this statistic in those versions and the
unreliability in attempting to add the numbers seen in V$SGASTAT to
sum to SHARED_POOL_SIZE...
Then, sometime in the Oracle8 or Oracle8i timeframe, the meaning of
the statistic was changed so that the term "free memory" became what
everyone had thought it was, an adjective and a noun (i.e. as in "free
beer" or "free time"). A much more useful statistic, certainly...
Is this true? If not, is it close?
The sum of the information in V$SGASTAT still does not add to
SHARED_POOL_SIZE, though (query from v8.1.7.4.0 shown below):
SQL> select name, bytes from v$sgastat
2 where pool = 'shared pool';
NAME BYTES
-------------------------- ----------
free memory 18208352
miscellaneous 2378964
DML locks 120000
PLS non-lib hp 2096
trigger inform 944
PL/SQL MPCODE 1146204
PL/SQL DIANA 1223360
PX subheap 123476
db_block_hash_buckets 1411080
sessions 377300
KGK heap 48124
State objects 267420
message pool freequeue 124552
Checkpoint queue 885168
enqueue_resources 222912
db_files 370988
KGFF heap 649844
KQLS heap 1709904
dictionary cache 12670280
table definiti 3228
transactions 171264
ksfv subheap 4248
fixed allocation callback 1280
library cache 89490788
simulator trace entries 240000
sql area 187432036
table columns 19520
processes 123380
partitioning d 152976
db_block_buffers 10880000
event statistics per sess 607600
----------
sum 331067288
SQL> show parameter shared_pool_size
NAME TYPE VALUE
------------------- ------- ---------
shared_pool_size string 314572800
I'm curious about the 16,494,488 bytes difference. Is it possible
that V$SGASTAT is another "unlatched" data structure in memory,
allowing errors in the interest of eliminating contention? There are
other similar structures in the SGA (i.e. the data structure
underlying table MONITORING statistics later flushed to
SYS.TABMOD$)...
Thanks for any and all insight!
-- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: =?ISO-8859-1?Q?Mogens_N=F8rgaard?= INET: [EMAIL PROTECTED]
Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California -- Mailing list and web hosting services --------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
