SGA/PGA (Memory) in RAC environment.
Guys,
I would like to get your inputs on SGA / PGA settings on a RAC database.. Does LMS/LMON/LMD process use memory which is part of PGA or these process allocate memory at OS level ?
I would like to see the usage of Memory used by a database.
below sql from Mr. Kyte's website i have used. Somehow i am not getting real usage of the system. Any help would be appreciated.
Thanks
select
ssst.sid,
stn.name ,
round(ssst.value/1024/1024,2) memory ,
USERNAME,
COMMAND,
OSUSER,
MACHINE,
TERMINAL,
PROGRAM ,
ssst.STATISTIC#
from v$statname stn,v$sesstat ssst , v$session ses
where stn.STATISTIC# = ssst.STATISTIC# and
ssst.sid = ses.sid and
name like 'session%pga%memory%'
order by 3 asc ;
Edited by: P explorer on Aug 28, 2009 9:56 PM
From sql script, that get "session pga memory","session pga memory max" each of session in oracle database(on that node)
You'll see
sid: session id
memory:
if "session pga memory" = pga memory in this session now
if "session pga memory max"= maximum pga memory in this session.
program: os program name
More deatails abour v$session http://download.oracle.com/docs/cd/B19306_01/server.102/b14237/dynviews_2088.htm#REFRN30223
If need to know all PGA size in that node
"session pga memory":
select
round(sum(ssst.value)/1024/1024,2) memory
from v$statname stn,v$sesstat ssst , v$session ses
where stn.STATISTIC# = ssst.STATISTIC# and
ssst.sid = ses.sid and
name ='session pga memory';
"session pga memory max":
select
round(sum(ssst.value)/1024/1024,2) memory
from v$statname stn,v$sesstat ssst , v$session ses
where stn.STATISTIC# = ssst.STATISTIC# and
ssst.sid = ses.sid and
name ='session pga memory max';
Or if you need to check PGA memory all nodes (RAC)
COLUMN username FORMAT A20
COLUMN module FORMAT A20
SELECT a.inst_id,
NVL(a.username,'(oracle)') AS username,
a.module,
a.program,
Trunc(b.value/1024) AS memory_kb
FROM gv$session a,
gv$sesstat b,
gv$statname c
WHERE a.sid = b.sid
AND a.inst_id = b.inst_id
AND b.statistic# = c.statistic#
AND b.inst_id = c.inst_id
AND c.name = 'session pga memory'
AND a.program IS NOT NULL
ORDER BY b.value DESC;
Similar Messages
-
Reg:-SGA & PGA Memory Allocation Size
Hi all
I Have a server System Configuration with
RAM size: 16GB Ram & Hard disk Size of 1TB.
I Installed Oracle 11g R2 in the server System
I want to Allocate Dedicated Memory for oracle..
How Should i calculate the Memory size of SGA+PGA for good Performance..Check MOS
*How to calculate the memory used by dedicated and shared servers? [ID 551415.1]* -
Hi all
I Have a server System Configuration with
RAM size: 16GB Ram & Hard disk Size of 1TB.
I Installed Oracle 11g R2 in the server System
I want to Allocate Dedicated Memory for oracle..
How Should i calculate the Memory size of SGA+PGA for good Performance..Pl do not post duplicates - Reg:-SGA & PGA Memory Allocation Size
-
Calculating total memory in oracle RAC environment
I have to calculate total memry in RAC environment.
For shared and buffer pool I execute show sga.
For UGA and PGA I execute statement that have two different values.
This is my two different methot for calculating total memory in oracle RAC environment.
Why I have very different value in this 2 statements on pga values?
first stat
with vs as
select 'PGA: ' pid
,iid
,session_pga_memory + session_uga_memory bytes
from (select inst_id iid
,(select ss.value
from gv$sesstat ss
where ss.sid = s.sid
and ss.inst_id = s.inst_id
and ss.statistic# = 20) session_pga_memory
,(select ss.value
from gv$sesstat ss
where ss.sid = s.sid
and ss.inst_id = s.inst_id
and ss.statistic# = 15) session_uga_memory
from gv$session s)
union all
select 'SGA: ' || name pid
,s.inst_id iid
,value bytes
from gv$sga s
select distinct iid, pid, sum(bytes) over (partition by iid, pid) bytes from vs
IID PID BYTES
1 PGA: 196764792 <=====
1 SGA: Database Buffers 318767104
1 SGA: Fixed Size 733688
1 SGA: Redo Buffers 811008
1 SGA: Variable Size 335544320
2 PGA: 77159560 <=====
2 SGA: Database Buffers 318767104
2 SGA: Fixed Size 733688
2 SGA: Redo Buffers 811008
2 SGA: Variable Size 335544320
second stat
with vs as
select 'PGA: ' pid
,p.inst_id iid
,p.pga_alloc_mem bytes
from gv$session s
,gv$sesstat pcur
,gv$process p
where pcur.statistic# in ( 20 -- = session pga memory
,15 -- = session uga memory
and s.paddr = p.addr
and pcur.sid = s.sid
and pcur.INST_ID = s.INST_ID
and pcur.INST_ID = p.INST_ID
union all
select 'SGA: ' || name pid
,s.inst_id iid
,value bytes
from gv$sga s
select distinct iid, pid, sum(bytes) over (partition by iid, pid) bytes from vs
IID PID BYTES
1 PGA: 342558636 <=====
1 SGA: Database Buffers 318767104
1 SGA: Fixed Size 733688
1 SGA: Redo Buffers 811008
1 SGA: Variable Size 335544320
2 PGA: 186091416 <=====
2 SGA: Database Buffers 318767104
2 SGA: Fixed Size 733688
2 SGA: Redo Buffers 811008
2 SGA: Variable Size 335544320I'm sorry but it is not clear to me.
- From v$session (1th stmt) I have
nearly 196MB of PGA mem on instance 1
and
nearly 77MB of PGA mem on instance 2
- From v$process (2th stmt) I have
nearly 342MB of PGA mem on instance 1
and
nearly 186MB of PGA mem on instance 2
then...
342+186 - 196+77 = nearly 255MB of memory allocated by oracle processes but free?
if I want calculate the total thing of the amount of the allocated memory from Oracle...It is more correct 2th statement that query v$process...it is true? -
PGA memory Leak issue along with TNS: listener could not hand off client connection
Hi,
We have two Oracle Database instances in a 32 bit Standard Edition Windows OS machine with 3 GB RAM : 1) 1525 MB (SGA+PGA) 2) 515 MB (SGA+PGA)
Recently we upgraded the RAM from 3 GB to 4 GB. Subsequently we upgraded the SGA + PGA of the 2nd instance to 771 MB.
This has resulted into PGA memory leak issue reported in the alert log file and trace file.
Reverting back the SGA+PGA size to 515 MB solved the issue.
Is this issue related to Address Windowing Extension (AWE)?
Is the reason related to Total memory for all the cumulative Oracle instance should be < 2GB.
How to solve this?
Mon Jul 15 08:21:25 2013
ORACLE V10.2.0.4.0 - Production vsnsta=0
vsnsql=14 vsnxtr=3
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Windows Server 2003 Version V5.2 Service Pack 2
CPU : 4 - type 586, 4 Physical Cores
Process Affinity : 0x00000000
Memory (Avail/Total): Ph:1068M/4023M, Ph+PgF:4985M/7933M, VA:18M/2047M
Instance name: dp3
Redo thread mounted by this instance: 0 <none>
Oracle process number: 0
Windows thread id: 4484, image: ORACLE.EXE (SHAD)
*** 2013-07-15 08:21:25.133
*** ACTION NAME:() 2013-07-15 08:21:25.133
*** MODULE NAME:() 2013-07-15 08:21:25.133
*** SERVICE NAME:() 2013-07-15 08:21:25.133
*** SESSION ID:(90.829) 2013-07-15 08:21:25.133
******** ERROR: PGA memory leak detected 6302928 > 30176 ********
HEAP DUMP heap name="pga heap" desc=6C98A800
extent sz=0x206c alt=92 het=32767 rec=0 flg=3 opc=3
parent=00000000 owner=00000000 nex=00000000 xsz=0x0
EXTENT 0 addr=7FC10008
Chunk 7fc10010 sz= 64996 free " "
Chunk 7fc1fdf4 sz= 2097676 freeable "krfrbAllocate: "
Dump of memory from 0x7FC1FDF4 to 0x7FE20000
7FC1FDF0 1020020D 7FC10010 03871E90 [.. .........]
7FC1FE00 00000000 00000000 00000000 00000000 [................]
Repeat 131103 times
EXTENT 1 addr=7FA00008
Chunk 7fa00010 sz= 64996 free " "
Chunk 7fa0fdf4 sz= 2097676 freeable "krfrbAllocate: "
Dump of memory from 0x7FA0FDF4 to 0x7FC10000
7FA0FDF0 1020020D 7FA00010 03871E90 [.. .........]
7FA0FE00 00000000 00000000 00000000 00000000 [................]
Repeat 131103 times
EXTENT 2 addr=7F7F0008
Chunk 7f7f0010 sz= 64996 free " "
Chunk 7f7ffdf4 sz= 2097676 freeable "krfrbAllocate: "
Dump of memory from 0x7F7FFDF4 to 0x7FA00000
7F7FFDF0 1020020D 7F7F0010 03871E90 [.. .........]
7F7FFE00 00000000 00000000 00000000 00000000 [................]
Repeat 131103 times
EXTENT 3 addr=70A50008There may be something fixed in the patch that gives the error rather than mysterious hangs. I would prefer that, myself. There may be something in the patch that changes how Oracle allocates user, system and program areas.
At some point you spend more trying to limp things along in a growing environment than simply upgrading. 32 bit environments these days is an indicator that management is figuring some false economies. But that's their prerogative.
The PGA can be affected by many things, not least of which is how the application uses it. Much has been written about it, my own experience has been more is used than most people say. But that is probably just my app. Note there are different ways to set it, and a parameter with "target" in it might not be a hard limit. Different execution plans for the same code may use different amounts of memory - there can even be situations where forcing some code to use a sort area instead, or even forcing more operations to disk, can free up more pga for everyone else. It depends. But that's why I look at multi-pass operations in dbconsole. -
SGA/PGA Sizes ???
I need help!
I have a Windows 2003 Server (w/sp2) with 4GB of RAM & 2 CPUs. Oracle 10.2.0.2 is installed.
ASMM is ENABLED by default (and it still is)
These are my sga/pga settings:
pga_aggregate_target = 1G
sga_max_size = 1504M
sga_target = 1000M
shared_pool_size = 504M
sort_area_size = 262144000 (250M)
large_pool_size = 104M
db_cache_size = 104M
I believe I should DISABLED ASMM, correct? Otherwise, my settings aren't used.
Based on my 4GB, I think Oracle can use 2GB of that.
What should I set the settings above too?
...thanksFirst if you use Oracle's automatic memory management feature and provide spfile parameter values for parameters that are automatically managed by Oracle then the provided values are taken as minimum values for these areas. There are other situations where values for parameters like sort_area_size which is limited to 5% of pga_aggregate_target are in fact still referenced.
With a pga_aggregate_target of 1G I do not see any valid reason why you would have a sort_area_size set to 250M. What is the reason for this?
What kind of environment is this: OLTP, DSS, OLAP?
How many concurrent sessions?
Dedicated vs shared server?
Why do you not want to use automatic memory management? What problem have you encountered?
HTH -- Mark D Powell -- -
PGA memory problem - Oracle 10.2.0.4 on windows 2003
Hi,
I have recently started work at a new company and we are running Oracle 10g (10.2.0.4, Enterprise Edition) on Windows 2003 (Standard Edition). The server has 4Gb of RAM (and we have modified boot.ini to inclue the /3Gb switch).
RE: SGA/PGA, we have the following Oracle parameters set:
sga_target 1G
pga_aggregate_target 194M
The employees tell me that they frequently "have to reboot the database" because of ORA-4030 and ORA-4031 problems. Looking at taskmgr on the server, Oracle is using "too much" memory (~3Gb). New sessions cannot connect etc. and they restart the database. Being a DBA (experience in UNIX, not Windows) I'm not so keen on this "solution" and am trying to find out what's happening.
When this problem occurred yesterday, before allowing the reboot, I bought myself some time to have a little dig around in the database. In v$sesstat I saw one process that had a value of over 1GB for "session pga memory". Memory usage on the server for oracle.exe was (as predicted) ~1GB over the "expected" 1.2Gb value (of SGA+PGA agg target). So, part 1 of my question is:
- Is this "normal" behaviour for Oracle to allow a process to go so wild on the PGA?
(I understood that Oracle would attempt to maintain total PGA memory close to the value of PGA_AGGREGATE_TARGET. I believe I read in the documentation that it could allow PGA memory to increase "up to 20% over this value" but please don't quote me on that, I can't find it again..)
Part 2 to this problem is that sessions "collect" in the database and do not release their PGA memory, leading to the slow build up of memory until the errors are encountered. I believe Dead Connection Detection (DCD) not to be working here (sql_net.expire_time=1 is set on the server but appears to do nothing). I've started reading docs/notes on this and it seems that DCD is not reliable on Windows. Metalink Doc 151972.1 suggests testing and adjusting some underlying TCP/IP serttings in the O/S kernel (I'm not even sure how to do this in Windows yet, let alone if it's something I want to get involved with!). So:
- I'm wondering if anyone has any tips for this (killing off dead connections, getting DCD working in Windows 2003 etc.)? Any experiences, tips welcome here!!
Thanks in advance.
Regards,
Adosuser647632 wrote:
(By the way, can anyone recommend how to sort the formatting of these results please?!!)You can find all by clicking the Oracle Forum FAQ
</br>
Here is my PGASTAT result. Have a look at the values.
SQL> column name format a60
column value format 9,999,999,999,999,999
select * from gv$pgastat order by inst_id,name;
SQL> SQL>
INST_ID NAME VALUE UNIT
1 PGA memory freed back to OS 202,362,322,944 bytes
1 aggregate PGA auto target 1,831,209,984 bytes
1 aggregate PGA target parameter 2,147,483,648 bytes
1 bytes processed 287,247,907,840 bytes
1 cache hit percentage 68 percent
1 extra bytes read/written 133,790,002,176 bytes
1 global memory bound 214,743,040 bytes
1 max processes count 48
1 maximum PGA allocated 1,708,733,440 bytes
1 maximum PGA used for auto workareas 1,112,871,936 bytes
1 maximum PGA used for manual workareas 271,360 bytes
INST_ID NAME VALUE UNIT
1 over allocation count 0
1 process count 42
1 recompute count (total) 136,756
1 total PGA allocated 328,158,208 bytes
1 total PGA inuse 196,502,528 bytes
1 total PGA used for auto workareas 81,608,704 bytes
1 total PGA used for manual workareas 0 bytes
1 total freeable PGA memory 96,927,744 bytes
19 rows selected.
SQL>
SQL> column BYTES_PROCESSED format 9,999,999,999,999,999
column EST_RW_EXTRA_BYTES format 9,999,999,999,999,999
SQL> SQL> select inst_id,round(pga_target_for_estimate/1024/1024) as target_size_MB,
2 bytes_processed,estd_extra_bytes_rw as est_rw_extra_bytes,
3 estd_pga_cache_hit_percentage as est_hit_pct,
4 estd_overalloc_count as est_overalloc
5 from gv$pga_target_advice order by inst_id,target_size_mb;
INST_ID TARGET_SIZE_MB BYTES_PROCESSED EST_RW_EXTRA_BYTES EST_HIT_PCT EST_OVERALLOC
1 256 285,418,388,480 188,648,610,816 60 4
1 512 285,418,388,480 131,006,145,536 69 0
1 1024 285,418,388,480 92,476,995,584 76 0
1 1536 285,418,388,480 91,536,565,248 76 0
1 2048 285,418,388,480 72,373,725,184 80 0
1 2458 285,418,388,480 68,650,139,648 81 0
1 2867 285,418,388,480 68,650,139,648 81 0
1 3277 285,418,388,480 68,650,139,648 81 0
1 3686 285,418,388,480 68,650,139,648 81 0
1 4096 285,418,388,480 68,650,139,648 81 0
1 6144 285,418,388,480 68,650,139,648 81 0
INST_ID TARGET_SIZE_MB BYTES_PROCESSED EST_RW_EXTRA_BYTES EST_HIT_PCT EST_OVERALLOC
1 8192 285,418,388,480 68,650,139,648 81 0
1 12288 285,418,388,480 68,650,139,648 81 0
1 16384 285,418,388,480 68,650,139,648 81 0
14 rows selected.
SQL>
SQL> show parameters pga
NAME TYPE VALUE
pga_aggregate_target big integer 2G
SQL> show parameters sga_max
NAME TYPE VALUE
sga_max_size big integer 2G
SQL> show parameters sga_target
NAME TYPE VALUE
sga_target big integer 2G
SQL> -
Hi I installed oracle 11g & create database , now i want to use fully optimize database.
Currently my DB server RAM is 16GB, how should i allocate SGA for this.If i used oracle 11g automatic memeory management feature to allocate SGA+PGA, how should i use them.Currentlymy SGA is set to 4GB, Please adviseHi,
It is depends upon the workload which you have in your organization(future and present situation)
SGA+PGA size will be automatically managed by Oracle. No need to worry about this.
If you need to increase Memory_Max_Size, you can also increase the Memory size in the future.
Thanks -
Oracle Streams on a Rac Environment
Hi
I have some questions with respect to Setting up Streams on a rac Environment.Would appreciate a quick response as I need the answers by tommorrow.Any help would be greatly appreciated.Here are the questions
1> Do we have to create capture process for each active instance on only 1 capture process will do?
2> If yes then do they need to have a seperate queue for each one?
3> How will the apply process access multiple capture process and the propogation take place?
4> can only 2 tables in the source be replicated instead of the entire database?
5> In case if we use a push job if both the primary and secondary go down how can we move to the third instance and use it?
6> If the instance goes down do we have to restart the capture process once again?
7>What is the best suited for rac - ASM/RAW FILES with respect to Streams?
Regards
ShwetaStreams in 9iR2 RAC environment mines only from archive logs not online redo logs. This restriction is lifted in 10g RAC. If you choose to go thru the downstream capture route in 10g then you can only mine from archive logs in 10gR1.
Having said the above here are my answers:
1> Do we have to create capture process for each active instance or only 1 capture process will do?
You can run multiple capture processes each on difference instance in RAC. Unless you have a requirement to do so, a single capture process would suffice. The in-memory queue should also be on the same instance as the capture process is running from.
2> If yes then do they need to have a seperate queue for each one?
YES
3> How will the apply process access multiple capture process and the propogation take place?
Propagation is from a source queue to the destination queue. If the destination is a single instance database, then you can direct propagations for all of your capture(s) into a single apply queue. If the destination is also RAC then you can run multiple apply processes on each node and apply changes for specific set of tables. Maintenance would be something to think about here along with what happens when one node goes down.
4> can only 2 tables in the source be replicated instead of the entire database?
YES. Streams is flexible to let you decide what level you want to replicate.
5> In case if we use a push job if both the primary and secondary go down how can we move to the third instance and use it?
In theory propagation is a push job. There are certain things you need to configure correctly. If done, then you can move the entire streams configuration to any of the surviving node(s).
6> If the instance goes down do we have to restart the capture process once again?
In 9iR2 you have to restart the streams processes. In 10g the streams processes automatically migrate and restart at the new "owning" instance. In both versions, Queue ownership is transferred automatically to the surviving instance.
7> What is the best suited for rac - ASM/RAW FILES with respect to Streams?
Streams is independent of the storage system you use. I cannot think of any correlation here. -
Hi
We're doing some performance testing on our database. The SGA is set to 1.5GB and PGA set to 0.5GB. At some point we are runnning out of memory as we ramp up the number of users and transactions .
What I'm looking for is to graph the PGA and SGA usage seperately regular intervals (a line graph) showing their increaase as users ramp up.
Anyone any ideas on how I can get this information whether through AWR or manually.
Thanks in advance.You have not posted the version you are on. You should use automatic pga management and unset the values for areasize parameters (if there is no specific reason for keeping them).
ORA-04030: out of process memory when trying to allocate 8389132 bytes (pga heap,redo read buffer)
Doing block recovery for file 3 block 76812
Resuming block recovery (PMON) for file 3 block 76812What are the OS level statistics of memory usage ?
It simply means that Oracle (rather OS...as Oracle is also going to request memory from OS only) is running out of memory. Most common reason behind that is use of features like bulk collect without any limit clause which causes the pga memory to grow beyond pga_aggregate_target...and finally sucking all the memory available. Further requests for memory allocation start failing with ORA-04030.
Also, please post the output from v$pgastat using tags.
Edited by: amardeep.sidhu on Nov 24, 2010 9:29 AM -
Hi all,
My questions is regarding how to set up SGA & PGA target values giving DB server is MS 2003 Enterprise 64 edition sever with 63 GB of RAM. Our DB’s version is 10.2.0.3 and RACed with 2 instances.
Currently our setting for SGA is 23 GB and PGA is 4 GB. I ran the memory advisor it display different optimal values for those two instances (35 GB and 23 GB for SGA, 5 GB and 4 GB for PGA). Guess our db is not well balance loaded.
I would like to know if there is any rule(s) that SGA can’t be more than certain percentage of total RAM for the server.
Thanks a lot in advance.
Have a great weekend!!
ShirleyHi
Here is my approach:
1) consider free physical memory (meassure must be taken with Oracle down)
2) substract a 10% of that amount as a margin
3) if the box is exclusive for Oracle, the remaining amount may be your SGA+PGA limit
4) if not is exclusive, then substract memory required for other services, then you'll get the SGA+PGA limit.
5) Substract your PGA size, giving the SGA figure
For instance
59 Gb - 6 Gb = 53 Gb - 4 Gb = 49 Gb
(63 Gb -4 Gb [estimated OS footprint + IO cache]) = 59 Gb free physical memory
Setting PGA+SGA beyond the free physical memory figure, may lead to swaping and degraded system performance.
Regards
http://oracledisect.blogspot.com -
SGA, PGA and... little extra?
Will oracle process use some extra "non oracle" memory for OS related stuff?
Mean, not considering SGA, I connect a database process and I see:
SQL> select p.program,
2 p.spid, pm.category, pm.allocated/1024 allocated, pm.used/1024 used, pm.max_allocated/1024 max_allocated
3 from V$PROCESS p, V$PROCESS_MEMORY pm
4 where p.pid = pm.pid
5 and p.spid =2650;
Max
Allocated Usedd allocated
PROGRAM SPID CATEGORY Kb Kb Kb
oracle@uxlabhost11 2650 SQL 14 9 42
oracle@uxlabhost11 2650 PL/SQL 56 34 57
oracle@uxlabhost11 2650 Other 1501 1501That's allocated PGA after having connected and done nothing else
So it somehow allocated 1.4 Mb upon start-up...
Now question is. Do oracle processes use OS memory which cannot be considered as SGA or PGA?
Reason being my trying to estimate memory needs for a database server. Many I've read correct formula should be SGA + max allocated PGA + some_extra_just_in_case...
Is there an estimate on how much that "some_extra_just_in_case" would be per process?
If OS dependent, I'm implementing on RHEL 6...
thanksSo, I've used pmap...
Got 257 enties in the form "4096K rw-s- /dev/shm/"
since we are using AMM , that should do for PGA and SGA ... some 1052672 Kb, which process is attached to...
yet pmap then continues to display other entries....
0000002a95557000 4K r-x-- /opt/oracle/product/11g/lib/libodm11.so
0000002a95558000 1024K ----- /opt/oracle/product/11g/lib/libodm11.so
0000002a95658000 4K rw--- /opt/oracle/product/11g/lib/libodm11.so
0000002a95659000 4K rw--- [ anon ]
0000002a9565a000 392K r-x-- /opt/oracle/product/11g/lib/libcell11.so
0000002a956bc000 1024K ----- /opt/oracle/product/11g/lib/libcell11.so
0000002a957bc000 40K rw--- /opt/oracle/product/11g/lib/libcell11.so
0000002a957c6000 4K rw--- [ anon ]
0000002a957c7000 864K r-x-- /opt/oracle/product/11g/lib/libskgxp11.so
.. TRUNCATED...
0000003fb810a000 64K rw--- [ anon ]
0000003fbbd00000 84K r-x-- /lib64/libnsl-2.3.4.so
0000003fbbd15000 1020K ----- /lib64/libnsl-2.3.4.so
0000003fbbe14000 4K r---- /lib64/libnsl-2.3.4.so
0000003fbbe15000 4K rw--- /lib64/libnsl-2.3.4.so
0000003fbbe16000 8K rw--- [ anon ]
0000007fbffde000 136K rwx-- [ stack ]
ffffffffff600000 4K r-x-- [ anon ]
total 1276452KNotice "total 1276452K"
So you saying non SGA/PGA would be 1276452 - 1052672?
That would be too much... yet if pmap another process, almost all flags (hex column) but [ stack ] are the same....
So how to tell which of those are shared and which are "exclusive" to the process, i.e: The "some extra just in case" mem...
thanks -
Hi all,
I using Database version 9.0.2.1 on Windows.
In Udump\.....trace file have a message:
*** 2004-04-11 13:28:58.000
*** SESSION ID:(4.1) 2004-04-11 13:28:57.000
******** ERROR: PGA memory leak detected 10340 > 6004 ********
HEAP DUMP heap name="pga heap" desc=001CDF10 ....
and then about 1-2 hours, my database is down (not connect) (error: ORA=3113: end of file communicational...)
I think, cause is PGA memory too large, may be ? I set value of PGA = 80-100 MB.
My server has 1.5GB RAM, (SGA = 900, Shared Mem=500, Buffer mem= 150..). It's correct ?
Please help me to resolve this problem.!!!!!
Thanks
[email protected]Hi ER,
It seems that you need to Update Patch - 9.2.0.6 which will solve this PGA Memory Leak!
Check this Doc ID: Note:283897.1 in Metalink!
Ravi Prakash -
Hi to All,
Exist a method in order to optimize SGA & PGA, installing a DB in Oracle DB 10g?
Best Regards
GaetanoAutomatic Shared Memory Management(ASMM)
Requirements:
SGA_TARGET >0
STATISTICS_LEVEL=TYPICAL or ALL.
The below mentioned parameters are auto-tuned by oracle.
SHARED POOL
LARGE POOL
JAVA POOL
DB CACHE
Refer to Note:295626.1 from Metalink.(SGA)
Refer to Note:223730.1 from metalink (PGA) -
Hi all;
i wanna learn something about sga and pga... I have r11 instance on linux and db version is 10.2.0.3...Let say my sga=5 Gb and pga is=3gb...my server has 20gb ram...when i look my usage by top command and i see that:
Mem: 20550320k total, 20491732k used why linux allocate all those area?
i have 20 Gb ram and sga+pga=8 soo which oracle process use 12 Gb?
and advicer gives me adivice about SGA... it offers 10 Gb SGA... if i rise my sga what is my benefit? what it would change?
thanks alotHi hussien;
Thanks for again answers and help
You need to identify what other processes are running on this node. For example, if you have CM running on the database tier node, then it is expected to find more memory allocated. You can simply stop all the services on the server and start it one by one, and each time and keep an eye on the allocated memory.Its my mistake, coz i talked wiht my client and they said me they are working on this machine and thatswhy i had this memory usage.. sory for this mistake again (but your post helped me alot,coz i asked them what they are doing for a now ,and they gave me that answer :),thanks for leading)
Since you are on 10g, I would suggest you also go through the following document.
Note: 295626.1 - How To Use Automatic Shared Memory Management (ASMM) In Oracle10g
https://metalink2.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=295626.1
If you still have any queries, please feel free to post it here.Thanks for note.. i am reading it and its soo usefull for me, thanks for sharing your knowledge here wiht us hussein..
you are an oracle angle :)
regards
Helios
Edited by: helios on 12.May.2009 06:58
Maybe you are looking for
-
How do I find if IMAP is enabled on my company's exchange server
I am thinking that IMAP is enabled on my company's exchange server. How ever, I cannot connect to the the exchange server using my iphone? Can some one enlighten me if they have got this working? telnet abc.xyz.com 143 Trying xx.yy.ww.zz... Connected
-
Specifying where Reminders and Notes are saved.
I can't get my head around where the agenda pulls it's data from and where notes should be saved and the same with reminders. I'd like reminders to be editable on my computer or emailed to me or, at least, somewhere where I will see them later. Notes
-
JComboBox - Bias With White Space
Is there a simple way to bias an editable JComboBox with a single space so that the text isn't pushed up against the left side of the box? This is a solution but only works on uneditable JComboBoxes due to using a Renderer and not an Editor. http://f
-
Hi, i have an abap where i make a update bseg set mansp = 'value' where cond.......... The update in bseg works well, BUT: the corresponding fields in BSID or BSAD are NOT updated. As bseg is a cluster-table i thought that the update in bsid and bsad
-
When getting new software it did something to my phone and now it wont work
i dowloaded 2.0 and then it restored my phone and then "We could not complete your iTunes store request an unknown error occured (-9838)" please help me