Confusion about Automatic Shared Memory Management

Hi,
Oracle Database 10g includes the Automatic Shared Memory Management feature which simplifies the SGA memory management significantly. To use Automatic Shared Memory Management, we have to set the SGA_TARGET initialization parameter to a nonzero value and the STATISTICS_LEVEL initialization parameter to TYPICAL or ALL.
Oracle Database 10g Rel. 2 documentation, in some places, says that:
If SGA_TARGET is specified, then the following FIVE memory pools are automatically sized:
* Buffer cache (DB_CACHE_SIZE)
* Shared pool (SHARED_POOL_SIZE)
* Large pool (LARGE_POOL_SIZE)
* Java pool (JAVA_POOL_SIZE)
* Streams pool (STREAMS_POOL_SIZE)
Ref.:
1. http://download-uk.oracle.com/docs/cd/B19306_01/server.102/b14237/initparams192.htm
2. http://download-uk.oracle.com/docs/cd/B19306_01/server.102/b14220/memory.htm
3. Oracle Database 10g: New Features for Administrators - Student Guide
But in some places I found the following:
If SGA_TARGET is specified, then the buffer cache (DB_CACHE_SIZE), Java pool (JAVA_POOL_SIZE), large pool (LARGE_POOL_SIZE), and shared pool (SHARED_POOL_SIZE) memory pools are automatically sized.
Here you can see that Streams Pool is not included in the automatically sized pools.
Ref.:
1. http://download-uk.oracle.com/docs/cd/B19306_01/server.102/b14211/build_db.htm#sthref252
Also, according to Oracle Press' Book "OCP Oracle Database 10g: New Features for Administrators Exam Guide:
Under Automatic Shared Memory Management, the database manages the
following FOUR major components of the SGA, also known as the auto-tuned SGA
parameters:
■ Buffer cache (DB_CACHE_SIZE)
■ Shared pool (SHARED_POOL_SIZE)
■ Large pool (LARGE_POOL_SIZE)
■ Java pool (JAVA_POOL_SIZE)
It is important to understand that even under Automatic Shared Memory
Management, you still need to configure any SGA component other than the four
auto-tuned components. Following are the manually sized components of the SGA:
■ Redo Log Buffer
■ The KEEP and RECYCLE buffer caches (if specified)
■ The nonstandard block size buffer caches (if specified)
■ The new Streams pool SGA component
■ The new Oracle Storage Management (OSM) buffer cache, which is meant
for the optional ASM instance
Now my question is "IS Streams Pool an auto-tuned SGA parameter?"
Thanks in advance.
--Khan.

Hi,
I would advise you to read Document I.D. Note:295626.1 on Oracle Metalink.
It states that
When enabled, it lets Oracle decide of the right size for some components of the SGA:
SHARED POOL
LARGE POOL
JAVA POOL
DB CACHE (using the DB_BLOCK_SIZE value)
The SGA_TARGET value will therefore define the memory size sharable between auto-tuned and manual parameters.
The manual parameters are:
DB_<KEEP/RECYCLE>CACHESIZE
DB_nK_CACHE_SIZE (non default block size)
LOG_BUFFER
FIXED SGA
STREAMS_POOL_SIZE
Adith

Similar Messages

Questions about db_keep_cache_size and Automatic Shared Memory Management

Hello all,
I'm coming upon a server that I'm needing to pin a table and some objects in, per the recommendations of an application support call.
Looking at the database, which is a 5 node RAC cluster (11gr2), I'm looking to see how things are laid out:
SQL> select name, value, value/1024/1024 value_MB from v$parameter
2 where name in ('db_cache_size','db_keep_cache_size','db_recycle_cache_size','shared_pool_size','sga_max_size');
NAME VALUE VALUE_MB
sga_max_size 1694498816 1616
shared_pool_size 0 0
db_cache_size 0 0
db_keep_cache_size 0 0
db_recycle_cache_siz 0 0
e
Looking at granularity level:
SQL> select granule_size/value from v$sga_dynamic_components, v$parameter where name = 'db_block_size' and component like 'KEEP%';
GRANULE_SIZE/VALUE
2048
Then....I looked, and I thought this instance was set up with Auto Shared Mem Mgmt....but I see that sga_target size is not set:
SQL> show parameter sga
NAME TYPE VALUE
lock_sga boolean FALSE
pre_page_sga boolean FALSE
sga_max_size big integer 1616M
sga_target big integer 0
So, I'm wondering first of all...would it be a good idea to switch to Automatic Shared Memory Management? If so, is this as simple as altering system set sga_target =...? Again, this is on a RAC system, is there a different way to do this than on a single instance?
If that isn't the way to go...let me continue with the table size, etc....
The table I need to pin is:
SQL> select sum (blocks) from all_tables where table_name = 'MYTABLE' and owner = 'MYOWNER';
SUM(BLOCKS)
4858
And block size is:
SQL> show parameter block_size
NAME TYPE VALUE
db_block_size integer 8192
So, the space I'll need in memory for pinning this is:
4858 * 8192 /1024/1024 = 37.95.......which is well below my granularity mark of 2048
So, would this be as easy as setting db_keep_cache_size = 2048 with an alter system call? Do I need to set db_cache_size first? What do I set that to?
Thanks in advance for any suggestions and links to info on this.
cayenne
Edited by: cayenne on Mar 27, 2013 10:14 AM
Edited by: cayenne on Mar 27, 2013 10:15 AM

JohnWatson wrote:
This is what you need,alter system set db_keep_cache_size=40M;I do not understand the arithmetic you do here,select granule_size/value from v$sga_dynamic_components, v$parameter where name = 'db_block_size' and component like 'KEEP%';it shows you the number of buffers per granule, which I would not think has any meaning.I'd been looking at some different sites studying this, and what I got from that, was that this granularity gave you the minimum you could set the db_keep_cache_size, that if you tried setting it below this value, it would be bumped up to it, and also, that each bump you gave the keep_cache, would be in increments of the granularity number....?
Thanks,
cayenne

Large SGA On Linux and Automatic Shared Memory Management problem

Hello
I use Oracle10gR2 in linux 32bit and I use http://www.oracle-base.com/articles/linux/LargeSGAOnLinux.php manual
for larger SGA it works fine but when I set sga_target parameter for using Automatic Shared Memory Management
I recieve this error
ERROR at line 1:
ORA-02097: parameter cannot be modified because specified value is invalid
ORA-00824: cannot set sga_target due to existing internal settings, see alert
log for more information
and in alert log it has been wrote
Cannot set sga_target with db_block_buffers set
my question is when using db_block_buffers can't use Automatic Shared Memory Management ?
Is any solution for using both Large SGA and Automatic Shared Memory Management ?
thanks
Edited by: TakhteJamshid on Feb 14, 2009 3:39 AM

TakhteJamshid wrote:
Do it means that when we use large SGA using Automatic Shared Memory Management is impossible ?Yes its true. An attempt to do so will result inthis,
>
ORA-00825: cannot set DB_BLOCK_BUFFERS if SGA_TARGET or MEMORY_TARGET is set
Cause: SGA_TARGET or MEMORY_TARGET set with DB_BLOCK_BUFFERS set.
Action: Do not set SGA_TARGET, MEMORY_TARGET or use new cache parameters, and do not use DB_BLOCK_BUFFERS which is an old cache parameter.>
HTH
Aman....

In 11g, How to Enable Automatic Shared Memory Management (ASMM)

hi experts,
I have a new 11.2 g database and I want to configure it to use ASMM.
To enable ASMM, should I assign a non-zero size to the Memory_Target parameter or the SGA_Target ? I have read conflicting statements.
Thanks, John

If you mean Automatic Memory Management, pl see
http://download.oracle.com/docs/cd/E11882_01/server.112/e10595/memory003.htm#ADMIN11011
http://download.oracle.com/docs/cd/E11882_01/server.112/e10897/instance.htm#ADMQS12039
HTH
Srini

Oracle 9i Automatic PGA Memory Management

Hello,
my team and me, we are facing difficulties to change the size of the PGA used by our server processes for HASH JOIN, SORT... operators,
here you can see the results of "select * from v$pgastat":
[pgastat dynamic view results|http://pastebin.com/m210314dc]
We have been increasing consecutively our pga_aggregate_target parameter from 1.7 Gb initially to 4Gb then at the end 6Gb, the value of "Global memory bound" and " aggregate pga auto target" on the link above are still equal to 0.
I have been reading threads on the forum and documentation see below, I understand how the global memory manager (CKPT) computest the sql memory target and then the global memory bound, as far as I understand I can only "play" on the pga_aggregate_target value in order to increase the size of our PGAs (I exclude to play with hidden parameters).
- Joze Senegacnik: Advanced Management of working areas in Oracle 9i/10g : http://tonguc.yilmaz.googlepages.com/JozeSenegacnik-PGAMemoryManagementvO.zip
- Dageville Benoit and Zait Mohamed: SQL memory management in oracle 9i
Here different information that could be usefull:
OS: solaris 10 (db running in a non global zone)
Arch: 64-bit sparcv9 kernel modules
Physical memory: 32 Gb (being shared between all non global zones)
Oracle version: 9.2.0.5 32bits
Values of init parameters and hidden parameters that could be relevant:
[init parameters|http://pastebin.com/m40340cf4]
[hidden parameters|http://pastebin.com/m50d74c53]
Maybe useful queries:
over work areas views, I use the following script:
[wa_analysis.sql|http://pastebin.com/d606ebd9b]
and the result of it:
[result of script wa_analysis.sql|http://pastebin.com/m5f49a2e5]

Joze Senegacnik wrote:
- either your sessions are using a lot of memory for storing variables like pl/sql arrays which is subtracted from automatic management: PGA_AGGREGATE_TARGET - (aggregated persistent area + a part of the run time area of all server processes)
- you are hitting a bug
- or maybe something elseI am really happy you come to this conclusion too, they are the same we made with my team and we have submitting to Oracle support via metalink SR 3-1216060641, we were asking if we hit the following bug (in note 1) or we leak about pl/sql or java... or else indeed,
note 1: PGA_AGGREGATE_TARGET Assigned Memory Is Left Unconsumed When Set High [ID 844542.1]
Joze Senegacnik wrote:
I would like to know:
1.) what were the values for global memory bound and autotarget immediately (or in short time) after the database restart or when you have increased them Just after the restart of the database and just after the change of P_A_T, we query v$pgastat immediately after and the value of global memory bound and auto target were equal to 0 byte,
2.) If you are able to change value of PGA_AGGREGATE_TARGET (P_A_T) to 10GB what happens with global memory bound and auto traget. They should be positive at least for a short time. As this is a dynamic parameter you can change it for a short time, run queries and set it back.We plan to do this tonight, we have an "heavy" ITIL change management procedures that allow us to make changes approved by change manager and only during night maintenance window on production system, I come back to you tomorrow. But we have been increasing from 1,7Gb to 4Gb to 6Gb, each time I have been querying v$sgastat in the next 2 mins and global memory bound and auto target were equal to 0 byte.
3.) Have you checked on the OS level how much memory are using server processes - do these numbers come along with what Oracle says. Not during problematic activities, meaning active work areas performing HASH-JOIN, SORT... operators,
unfortunately it is a production system, even if he performs poorly, we are not allowed to try or retry the poor queries, but if it comes again I'll do it,
during low activities, here the results paste with the scripts I used:
[pga processes info in oracle|http://pastebin.com/f2e540062]
I spooled the result rows of this previous script in /var/tmp/pga_processes.log then I loop over all processes pid and display pmap output anon info like this:
h5. cat /var/tmp/pga_processes.log | awk -F' ' '{print $5}' | xargs -n 1 -i pmap -x {}| grep -v 'Addres' |egrep 'Kb' 2>&1 > /var/tmp/pga_processes_os.log
then I merge line by line the two files with unix paste command, here the results:
[os and oracle pga informations|http://pastebin.com/f4135c8a6]
4.) How many server processes are running on you system in average/max and are you using just dedicated processes or also shared?in average 250, we are only using dedicated processes,
5.) At time of low activity is the global memory bound still 0 or becomes > 0. I have been querying every 15 min during more than 24 hours low activities, it still stay to 0,
5.) Are you experiencing paging/swapping on OS level?No, here orca figures for details:
[free memory|http://img509.imageshack.us/img509/5897/ohuron1asd2gauge1024xfr.png]
swap
[pagein pageout|http://img121.imageshack.us/img121/6946/ohuron1asd2gaugepginper.png]
[memory usage|http://img19.imageshack.us/img19/2213/ohuron1asd2gaugeppkerne.png]
6.) Please post the result of: select * from X$QESMMSGA ;during low activities, [results X$QESMMSGA|http://pastebin.com/f61df7093]
While you will be answering to my questions I'll try to figure out what we can do to properly diagnose the problem. As you are on 9i it is a little bit harder.I am really kind of your help, as we say in my country, "if you need tow arms one day to carry something, call me."
--Jeremy Baumont

Confused about photos/shared/albums folder

Hi there,
I have an iPhone 5 iOS 8. under my photos: I have the collection of photos taken to date. under the Shared tab (iCloud photo sharing), I have activity, and another album; under the Albums tab, I have camera roll, my photo stream, videos, events from my mac, faces from my mac, recently deleted, and Last 12 Months from my mac.
I do not have enough storage in my iCloud to update. I would like to understand why I cannot delete photos in various albums that are "from my mac". In the Last 12 Months folder from My Mac, there are 320 photos. I believe these are stored on my macbook so I don't need all of those on my phone. How do I remove some of these photos that are taking up iCloud space?
why can't the photos under the Photos tab where collections is found be deleted? I am so confused. It seems like I have multiple copies of the same photo in different folders. Can someone set me straight on what is what. I can't seem to learn about the difference in any of the mac support resources.

Photo Library contains all of your photos except Camera Roll. Because you have only one album with photos (the other is empty) the number of Photo Library equals to "for itunes" album.
846 (camera roll)+788 (Photo Library) are all of your photos that you have in iPhone.
Everything else is a "shortcut" ("for itunes" folder points to Photo Library, "for screen" album may points to Camera Roll or/and Photo Library).
That means that you have only 1634 photos in your iPhone instead of 2536 that iTunes say.
Transfer your Camera Roll to your computer to sync it back later as an album. (like "for itunes").
Empty your Camera Roll (you have to do it from time to time anyway) and un-sync the "for itunes" album and everything else. Try to empty your iPhone from any photo in Photo app.
Check your storage in Settings>General>Usage>Photos and Camera. Slide to delete if it needs.
If everything is correct sync back the "for itunes" and the folder that contains your 846 photos from Camera Roll.
Are there also Videos in your iPhone?
When you sync the "for itunes" folder did you checked the "Include videos" (or something like that) in Photos tab of iTunes sync page?

Automatic PGA Memory Management in 9i and 10g

Hi,
I read the Doc# 223730.1 for Auto PGA advisory.
At one point in the document, there is given a query to calculate percentage of times work areas were executed in optimal, one-pass, or multi-pass mode since start-up.
My question is why there is a clause low_optimal_size > 64*1024); ---- for 64 K optimal size
Why it is safe to ignore all the workareas below 64 K.
How to decide which entries should be ignored?
Thanks,
Neeraj

I think 64K value used in that document is just an example.
You can use the prior query (one above which you have cited here) for finding non-empty buckets and then view histograms for range of your concern (the range which is getting used more frequently in your environment).

SQL memory manager latch at top of the list

select * from V$LATCH order by wait_time desc shows wait time = 451524661079 on SQL memory manager latch. What is making this so high?

user498912 wrote:
Thank you .. the pointer to the bug is helpful. We are having performance issues and complaints of slowness only during about a 2 or 3 hour window each weekday from about 7AM to 10AM EDT. This is definitely peak load. AWR is showing mostly disk i/o issues. Plan is to add about 20 GB more memory on the server. SGA currently at around 61GB with Automatic Shared Memory Management enabled. PGA advice showing nothing with a Cache Hit Percentage (%) 95.93
I would look at v$sga_resize_ops (or v$memory_resize_ops) to see if there is pressure to move memory around in that time period. Your symptoms could simply be showing load on the buffer cache causing the library cache to shrink leading to agressive demands for library cache locks etc. If so, then fixing a minum shared_pool_size (or generally going to manual memory management) may be the best bet. Also worth checking if there is some inefficient SQL running at that time that results in lots of random I/O - eliminating the cause of disk reads may be the best solution to reducing demand for memory.
Regards
Jonathan Lewis

Sort Area Size in Automatic memory management

Hello All
I am aware that the AREASIZE is ignored of the PGA_AGGREGATE_TARGET is set.
So how is it possible that if we incrase the SORT_AREA_SIZE, the performance improves?
does this necessarily mean that the PGA_AGGREGATE_TARGET was not set to a proper value that it instead used the SORT_AREA_SIZE instead?
thanks

Hi,
If you have set workarea_size_policy=auto then under the automatic PGA memory management mode, sizing of work
areas for all sessions becomes automatic and the AREASIZE parameters are
ignored by all sessions running in that mode.
In auto mode if you change any AREASIZE parameters will be ignored.
If you want to manually handle ARASIZE then turnoff the Automatic pga memory by setting workarea_size_policy=MANUAL and then your changes to parameter will take effect but it's advisable to set pga to automatic.
To check whether your pga is set proper or not check v$pga_target_advice view
SELECT round(PGA_TARGET_FOR_ESTIMATE/1024/1024) target_mb,
ESTD_PGA_CACHE_HIT_PERCENTAGE cache_hit_perc,
ESTD_OVERALLOC_COUNT
FROM V$PGA_TARGET_ADVICE;
This will give you how your pga is set.
chirag

How will i increase the shared memory of the graphic card of macbook pro 13-inch, early 2011 from 4gb memory?

how will i increase the shared memory of the graphic card of macbook pro 13-inch, early 2011 from 4gb memory?

The current version of Mavericks is 10.9.4. To update yours download and install OS X Mavericks 10.9.4 Update (Combo).
You can upgrade RAM to the maximum your model will support. That would be:
Maximum Memory
16 GB (Actual) 8 GB (Apple)
Memory Slots
2 - 204-pin PC-10600 (1333 MHz) DDR3 SO-DIMM
About OS X Memory Management and Usage
Using Activity Monitor to read System Memory & determine how much RAM is used
OS X Mavericks- About Activity Monitor
Adding RAM only makes it possible to run more programs concurrently. It doesn't speed up the computer nor make games run faster. What it can do is prevent the system from having to use disk-based VM when it runs out of RAM because you are trying to run too many applications concurrently or using applications that are extremely RAM dependent. It will improve the performance of applications that run mostly in RAM or when loading programs.

My MacBook pro 2012 Mountain Loin 8.2 don't clear the inactive memory automatically when memory needed for other program so I get page outs.

My MacBook Pro 2012 Mountain Loin 8.2 doesn't clear the inactive memory automatically when memory needed for other program so I get page outs. There are much inactive memory when page outs. Why my mac does this?

That's completely normal. If other programs require memory then the system will reallocate Inactive RAM as needed. Page Outs don't mean anything.
About OS X Memory Management and Usage
Reading system memory usage in Activity Monitor
Memory Management in Mac OS X
Performance Guidelines- Memory Management in Mac OS X
A detailed look at memory usage in OS X
Understanding top output in the Terminal
The amount of available RAM for applications is the sum of Free RAM and Inactive RAM. This will change as applications are opened and closed or change from active to inactive status. The Swap figure represents an estimate of the total amount of swap space required for VM if used, but does not necessarily indicate the actual size of the existing swap file. If you are really in need of more RAM that would be indicated by how frequently the system uses VM. If you open the Terminal and run the top command at the prompt you will find information reported on Pageins () and Pageouts (). Pageouts () is the important figure. If the value in the parentheses is 0 (zero) then OS X is not making instantaneous use of VM which means you have adequate physical RAM for the system with the applications you have loaded. If the figure in parentheses is running positive and your hard drive is constantly being used (thrashing) then you need more physical RAM.
Adding RAM only makes it possible to run more programs concurrently. It doesn't speed up the computer nor make games run faster. What it can do is prevent the system from having to use disk-based VM when it runs out of RAM because you are trying to run too many applications concurrently or using applications that are extremely RAM dependent. It will improve the performance of applications that run mostly in RAM or when loading programs.

Unable to start DSC Shared Memory

I am running LabView 7.0 (I know its an older setup), with DSC and Lookout OPC to do Serial communications to an Omron PLC. I needed to update the Ni-DAQ because the PCI-6220 board needs the driver. I upgraded to NI-DAQ 7.5 and now a get the Unable to start the DSC Shared Memory Manager Service. I have uninstalled the DSC and reinstalled it with no change. I found in the discussion boards that I need nidscsm.exe in the windows\system32 folder but it is not there. I am running Windows XP.
I have 2 setups like this. 1 PC still connects to the PLC even though the error pops up, another PC will not connect at all, lights blink but then comms timeout error.
Thanks!
Solved!
Go to Solution.

NI DAQ is now out of date, if possible you should use DAQmx instead. According to this chart, you can use any version before DAQmx 8.1. This new driver should work better with DSC. If you need some help converting your code check out this tutorial:Transition from Traditional NI-DAQ to NI-DAQmx in LabVIEW

Problems with memory management

Since my adoption of 10.7, I have noticed a problem that is really reminiscent of 10.2 and 10.3 of poor memory management. This is quite distressing as it has taken away from the performance of the operating system and has created many pages of VM files that are just below the engineering standards of Apple. Any suggestions along this line to clear this up would be greatly helpful. I have 3 GB of memory that is over utilized and over taxed.

I have read a bit about the "better" memory management, but that is all in theory. In the real world, Lion eats RAM and i upgraded from 4 to 8 Gb last week, that makes all the difference. If i close all apps now, 5.5 Gb RAM is still used, and i know that has something to do with Lions special workmemory management that is supposed to work to our favour, but it doesnt! Now after upgrading, fairly cheap but nothing for mainstream people, things work the way i suppose they are meant to work, but when Windows 7 and soon 8 are working towards less RAM needed, Apple seems to say one thing and do another. That is, they claim to have a improved (superior, fantastic?) system for RAM, but in reality it eats RAM for breakfast. Now i dont haveto close one app before opening another just to avoid the little spinning ball of colours.
Also Safari can use gigabytes of RAM on its own....not good.

Mountain Lion memory management Issue !!!

Hi,
I have a MacPro Mid-2010 quad-core intel xeon with 10 gb ram.
Im an audio engineer n i use this machine extensively for my work.
I recently updated to mountain lion.
Since updating the memory management has gone from amazing (never ran out of memory even while doing extensive tasks) to crap (running out of memory all the time).
At the moment im adding a lot of tracks to itunes while simultaneously using mp3 scan+repair to fix my mp3 files.
this to me doesnt sound like anything that my MacPro should have difficulty with but its sluggish. i can still type this message easily while the other 2 tasks are on but its just feeling slow.
now when i look at istat this is what i see:
now why would there be so much inactive memory n free mem only like 10 mb with 260 page outs?
i shouldnt be getting page outs with so much inactive memory.
getting page outs means buying more memory but i dont think i need it.
im unclear about mountain lions memory management.
please advise.
Thanks.
Gautam.

I'm using version 7.0.15104.778994 and having no trouble under Mountain Lion. I think that it was a July 10th release (coinciding with the GM developer release of ML).
What version do you have?
Clinton

Question about 11gR2 Grid, RAC, /dev/shm and Automatic Memory Management

Hello,
i've recently installed grid and rdbms software 11.2.0.2 on a two node Oracle Linux cluster with 128gb ram each node.
I'm using ASM to store data and ocr and I'm testing Automatic Memory Management.
When I finished Grid+RDBMS installation I've seen that /dev/shm size is 64gb (half of my total RAM).
I've created a database with dbca and when I was asked to choose if I wanted to use AMM I've noticed that I could
allocate only about 60gb for Oracle. If I chose more than 90gb I got an error saying:
Using Automatic Memory Management requires 60gb available in my two nodes.
The current available space in the two nodes is only 30gb and 30gb.
If you want to use AMM you should either free up some space in /dev/shm
or reduce the memory allocated to Oracle
I was wondering when (during the installation or the settings of kernel parameters) did I define the space of /dev/shm ?
Since I have 128gb of RAM wouldn't it be better to use more than 64gb of ram for my /dev/shm tmpfs partition ?
Is there a limit or a ratio for best practice for my RAM and the /dev/shm ?
thanks in advance.

user9051299 wrote:
Is the "half of the RAM size" a kernel's default value or Oracle's ? Neither. There are a number of unique factors that determine the best memory size and fit for Oracle - including just how much memory is effectively available (i.e. how much is needed for other services and processes).
And from what I understand i don't "break" any Oracle's best practice by increasing the /dev/shm right ?Correct. (at least none that I'm aware of, and none that I have read in Oracle's RAC Starter Kit documentation).

Confusion about Automatic Shared Memory Management

Similar Messages

Maybe you are looking for