OEM Instance activity / physical IO / redo size

Similar Questions

• REDO size interesting 10g, 11g - PLSQL loop Vs ETG

  Hello
  
  While doing a test recently, I had a very interesting observation what size roll FORWARD is.
  I have supplemental logging enabled on my 10g and 11g databases
  My log on the databases file size is 50 MB.
  
  
  31%
  ----------
  
  SQL > select SUPPLEMENTAL_LOG_DATA_MIN, SUPPLEMENTAL_LOG_DATA_PK,.
  2 SUPPLEMENTAL_LOG_DATA_UI of the v database $;
  
  SUBSECTIONS SUP SUP
  -------- --- ---
  YES IMPLIED NO
  
  30%
  ------
  
  SQL > select SUPPLEMENTAL_LOG_DATA_MIN, SUPPLEMENTAL_LOG_DATA_PK,.
  2 SUPPLEMENTAL_LOG_DATA_UI of the v database $;
  
  SUBSECTIONS SUP SUP
  -------- --- ---
  IMPLICIT YES YES
  
  
  Test 1:
  -------------
  
  declare
  n number;
  Start
  because me in 1... 2000000 loop
  n: = i;
  insert into test2 values (i);
  commit;
  end loop;
  end;
  /
  
  the end of it, here are my observations
  
  
  Table size = 26MB
  REDO 11 G ~ 2.5 GB (confirmed well AWR - redo size 2,045,222,732)
  REDO 10 G ~ 1 GB (confirmed well AWR - redo size 1,105,245,940)
  
  
  
  Test 2
  -------------------
  
  create table test3 in select * from test2;
  
  
  the end of it, here are my observations
  
  
  Table size = 26MB
  REDO 11G ~ 30 M
  REDO 10 G ~ 30 M (AWR confirmed)
  
  
  
  Now I find myself with so may questions
  
  (1) in the case where we ended a 26 MB of data insertion why there are heavy Generateion REDO (2.5 G 11 g! and 1 in 10!)
  (2) why is there much REDO difference between 10 G and 11 G (only difference is 11 G - 10 G Noarchive - Archive, but which does not do what a lot of difference)
  (3) in second why things are as expected even if it has same activity 26 MB
  
  
  I know there will be information from header etc... again, but that will be the same with both
  I'm surprised by so much geneartion again heavy
  
  How AGAIN is created and STORED in the case 1 and case 2
  
  Very interesting, but have no answers right now
  
  Appreciate your help
  
  Thank you
  Pramod

  894365 wrote:
  Why ETG generated very very less compared to the insertion of recovery. In the case of DEC is it going to be the only entry to REMAKE (more entries of space, the Index entries + additional entries management)? where it store information basis for this transaction (YVERT, RBA, blocks etc.) and the original data?

  Go back to stats and check the number of entries for the DEC recovery, it will be much more than one because of all the management space.

  DEC has write a vector of change again by block of data only and should not generate undo - which means that the total amount of recovery will be similar to the size of the table. There will be an excess because of the quntity of changes associated with the management of the space (when blocks are added to the object of the segment header block updates etc.). In fact, if you are running in noarchivelog mode, then DEC was default nologging - so you could create the table with a very small amount of logging. If you create a table with a primary key, there is a special optimization for the construction of the primary key, which means that it again generated for the index also only slightly larger than the size of the index.

  Concerning
  Jonathan Leiws

• Windows 7 OEM never active key but can't find discs, I can sell it?

  I have a spare Windows 7 OEM key, I bought several years for construction of the computer, but ended up never activate. I kept it in a journal of all keys that, over time, I bought, but I don't have the disks. MS has retained this page to download Windows 7 when you plug the key, I have confirmed here that the key is always good. I have an old SSD to sell with it, too. So it is possible to sell safe key with the SSD because it has never been activated?

  Yes, you can, as long as this is not enabled.

• Windows 7 OEM re-activation with re-replacement HD

  My current hard drive is in strange and I would replace it temporarily all trying to fix it.  I have two questions:

  (1) I'll be able to activate my Windows 7 OEM new temporary hard disk?

  and especially:

  (2) if I am able to activate Windows on the new drive, I'll be able to put my original hard drive, once it is fixed, in my computer and "re-install" the activation code on this HD previously enabled?   I no longer need it on the temporary disk.

  Thank you
  Karen

  There is not need - and no point - to 'disable' Windows before installing a new drive (or a new machine)

  Such a procedure is relevant that when you change the product key on the current installation, as no changes only occur on the Server Activation of Ms. MS senior officials also suggest that the procedure is actually dangerous!

  Unless you use recovery media from the manufacturer for the new installation, you MUST use the product key of the COA sticker or your full license to install retail package - if you try to use the key found a tool Keyfinder, chances are that it will not work.

• Windows 7 OEM re-activation after upgrading RAM (-2G)

  Hello

  I have a computer laptop lenovo, and a cleaning set up using the OEM through the DVD license from digital river. (Win 7 Home premium)
  Use it for awhile, I upgraded my RAM (-2G), then I'll have an invalid activation message.
  I thought this only happens when I change my motherboard, but sometimes also the RAM upgrade (+ 2G)?
  I heard that I had to call someone to fix it, but can't find the right number.
  Could someone help me on the number? or another way to solve this problem?
  Best,
  HW

  How to activate Windows 7 manually (activate by phone)
   
  1) click Start and in the search for box type: slui.exe 4
   
  (2) press the ENTER"" key.
   
  (3) select your "country" in the list.
   
  (4) choose the option "activate phone".
   
  (5) stay on the phone (do not select/press all options) and wait for a person to help you with the activation.
   
  (6) explain your problem clearly to the support person.
   
  http://support.Microsoft.com/kb/950929/en-us

• No way to force a re-an OEM SLP activation online a license for laptop?

  So I try to get my hands on the promo code of Windows $8 15 that my laptop is eligible. When I go through the form it tells me that I am not eligible based on my key. So I went through the customer service who finally told me that my Windows 7 key was not enabled. The key turned on of course, through the MSI OEM SLP key, so he has never activated online and the activation servers have obviously no record activation. Is it possible to force a reactivation online in Windows 7? I tried to run slui.exe 5 but he says could not connect to the activation servers. I tried to explain this to technical support but now I'm just ignored...

  Any help or clues would be greatly appreciated!

  Try to use

  SLUI 3

  instead of 5 SLUI

• 32-bit Windows 8 Pro OEM activation using a 64-bit OEM key activation - is it possible?

  I have one still to be put in service oem 64-bit of windows Pro 8 copy and I was eager to install the 32-bit version on an old pc running Windows XP and with only 4 GB of memory. The 64-bit activation key can be used for this installation of 32-bit version?

  I downloaded an iso of the equivalent of 32 bits, it installed OK but when I tried to activate it using the 64-bit activation key I got the message "this product key does not work. Is it because the key provided with the 64-bit oem disk is usable only on a 64-bit install (unlike the commercial versions non - oem where the same key works for 64-bit and 32 - bit)- or is the invalid key for some other reason?

  64-bit disc is titled HRM_CCSA_X64FRE_EN-US_DV5 and the 32-bit version, I downloaded HRM_CCSA_X86FRE_EN-US_DV5

  Product keys are not specific architecture (32-bit or 64-bit), the problem is, some keys do not work in some channels.

  A touch of detail will not work with OEM media and vice versa OEM key will not work with the media to detail.

• OEM 12 c - modification of Configuration - size medium to large size

  Hi all.

  When we installed the OEM 12 c here in the company, we looked at the possibility of installing the Medium Size (size of OEM deployment). But examining the hardware configuration and we are able to see this material can support the large size, this means 10,000 or several targets and 1000 or more agents, like the bellows of docs say. Is it possible to 'upgrade' to this size of deployment without having to install the software again?

  References:

  http://docs.Oracle.com/CD/E24628_01/install.121/e22624/preinstall_req_hw.htm#BACDDAAC

  http://docs.Oracle.com/CD/E24628_01/install.121/e24089/sizing.htm#BCGHIJEC

  Since you have already deployed EM, by changing the deployment type for med to large post install can be done by running a SQL in the database of rep that will set all the required parameters rep a large type of deployment. You can also do it manually, but I suggest this sql running.

  Download the large DB model type of deployment for the necessary platform for 12.1.0.3 EM - http://www.oracle.com/technetwork/oem/enterprise-manager/downloads/db-templates-1959276.html

  Unzip it and you will see this sql: set_repo_param_11.2.0.3_Database_SQL_for_EM12c_Large_deployment.sql

  Pls run it on your repository database and you are good to go now.

  For any other performance tuning see the doc you posted: Dimensioning of the Enterprise Manager deployment

• Estimating redo size of alter table DROP COLUMNS CONTINUE checkpoint of 10000

  Hello
  in my 10.2.0.3, I've got table size 90 GB suffering:
  ORA-12986: columns in partially suspended state. Propose to ALTER TABLE DROP COLUMNS CONTINUE
  I need to know how to estimate the size of redo of the:
  change the control of MORE of COLUMNS DROP table 10000 point.
  Is 2 x good estimate of table size?
  Kind regards.
  Greg

  GregG says:
  Please clarify for me how are distinguished between nitrifiants delay block trial and lower real column continue the process.
  Looks like DBC process will increment statistics such as:

  drain plug - number of calls ktugct
  gullies only - consistent read gets

  It's a good indicator, even if I think the specific statistics could be:

      commit txn count during cleanout
  

  I think it's one that tells you that you have made a change of "validation of transaction" to a block.

  I don't know why the CBD process is such redo intensive, is not on the update of ITL bit in db blocks only?

  It updates the inputs ITL and lock bytes, and it takes only entry of redo block; the overload of the roll forward database is in the tens of bytes is small compared to the 240 bytes per row of the column to fall - but if you have a large number of blocks to go before continue it really starts to have an effect, you might be fooled.

  WARNING - I haven't checked to see if delay block drain plug behaves (or even happens) on Delete column in the way that it would perform a simple update in bulk.

  Concerning
  Jonathan Lewis

• On RAC primary DB and a single Instance as physical Standby...

  It's my 11g Data Guard scenario:
  
  BECAUSE primary database 11g.
  
  Single Instance Oracle 11g.
  
  My first question is:
  
  If each node of the RAC has 3 online redo log, should I create 4 or 7 journal watch over the database pending?
  
  If restore database Standby long after the day I do full backup and I've got ti registry manually archived newspapers, what I do for a node of the RAC archived logs of the two?
  
  Thank you very much.

  In form 7
  You must register for both nodes. Remember that you have a LGWR instance, so a redolog thread instance and the two wires together are all changes to the database.
  Please make sure that in this configuration of the archived redo logs are on a shared disk.

  ----------
  Sybrand Bakker
  Senior Oracle DBA

• Windows 7 OEM itself activated after update

  My HP laptop was purchased in September 2010 and I did a windows update yesterday and its lost its settings in windows 7.  Wants a code for activation but the sticker dissipated.  If I contact someone with prrof of purchase to find the code?  I have all the packages and original invoices.

  Try to do a restore of the system before you start, facing this problem.

  How do I Do a System restore in Windows 7 or Windows Vista ...

  Option 1:

  If the COA sticker is damaged, you must use the recovery partition to reinstall Windows 7, it will be exempt from the need to use the product key to activate your installation of Windows 7.

  Option 2:

  If your recovery partition is not available or damaged, you must contact the manufacturer of your computer, and request that a defined recovery disk you can use to reinstall Windows 7. They could practice a small shipping and handling. This will exempt the need to also use the product key.

  Option 3:

  If neither options 1 or 2 is possible, then you will need to invest in a new license of full version of detail.

  New computers coming often pre-installed with Windows 7 have what is called a recovery partition. This is used to reinstall an operating system in the case of a system failure. To access it, you need to start when you start your computer by pressing a function key. This can be either F1, F2, F9, F10, F11, F12 key DEL or tab. See the manual that came with your PC for instructions on how to reinstall Windows.

  This is how the recovery partition is accessible to most popular brands...

  For HP, press F11 directly after switching on the device

• Windows OEM license for physical relationship.

  I ask this question in the forum Hardware because what I read; a Windows OEM license is married to a motherboard or hard drive.

  Is this true? The results of the research gave me conflicting answers.

  What happens if the motherboard or the C: drive dies? the associated Windows license would be lost with her?

  A computer components store told me that for the OEM depends more on what are the elements of qualification you purchase, or when you buy them. Which is in conflict with what I read.

  Note - my choice of OS Windows is Windows 7 Professional 64-bit, SP1, OEM.

  OEM qualifies for all DIY systems.

• scan_listener running on the node without instance active

  I have a 5 cluster nodes, we are runinng case ONLY nodes on 1,2,3 other INSTANCES (4 node and node 5) 2 are shutdown. (For the future, use the APPROPRIATE CASE). My scan_listeners are as follows:
  LISTENER_SCAN1 listener to SCAN is enabled
  SCAN of the earpiece that listener_scan1 is running on the node btpu15
  LISTENER_SCAN2 listener to SCAN is enabled
  SCAN of the earpiece that listener_scan2 is running on the node btpu13
  LISTENER_SCAN3 listener to SCAN is enabled
  SCAN of the earpiece that listener_scan3 is running on the node btpu12
  
  However, the instances work only on btpu11, btpu12 and btpu13.
  
  Also, when connecting to the database using SCAN, I do NOT connect to instance 1 which is located in btpu11. Connection, disconnection, reconnection, etc will bounce me instance 2 and 3, but never get to instantiate 1.
  
  Any help?

  Hello
  first of all it doesn't matter, where Auditors SCAN run in a CCR environment.
  They only send connections for local listeners, where database instances are running. It is permitted to have a listener to SCAN on a host without a database.
  Read this white paper for a clarification of the SCAN:
  http://www.Oracle.com/technetwork/database/clustering/overview/scan-129069.PDF

  Now that your first instance does not get connections, could have several questions:
  (a.) the database on node 1 is not reachable.
  => Please check with SCAN listeners, that the service registered for the database really also points to node 1 (service of lsnrctl LISTENER_SCAN1 on the SCAN respecitive listener node runs).
  => Please check with the local listening on node 1, that the listener is rising and services service (the GI Home lsnrctl service)
  (b) try to connect to the service directly on the local node (using the VIP (not SCAN)): sqlplus user/pwd @: 1521.
  -Online so you can connect and that you are on the instance on node 1, then maybe loadbalancing a kicking, and the cluster think there are too many sessions/load on server 1 to lead more connections to it.
  Try temporarily stopping the service on the node 2 and 3 and see if you get now a connections to server 1.
  If it works, then you can be sure loadbalancing took the decision to not connect to node 1...
  => If you are unable to connect, you have a configuration problem.

  On one of my groups I also had to make 20 to connect for a connection to a specific node, so do not count on it...

  Sebastian
  Sebastian

• 30EA2 - SQL Developer Vs Index SQL/Plus functional on column Date

  I just spotted this while using EA3.2, but it may also exist on other versions, I will consider. I'm on a 9i 9.2.0.7 box, but will see if I can recreate on an instance of 10 or 11
  
  The question is, if I create a shadow table
  
  CREATE TABLE paul_test
  (
  CREATION_DATE DATE,
  DATE of creation_date2,
  last_updsate_date DATE,
  date of last_updsate_date2
  )
  ;
  
  If I then create a functional index on creation_date via SQL Dev:
  
  create index pt_idx1 on paul_test (nvl (creation_date, ' 01-JAN-4000'));
  
  and create a functional index on creation_date2 via SQL more
  
  create index pt_idx2 on paul_test (nvl (creation_date2, ' 01-JAN-4000'));
  
  What I see when you perform a describe table statement and looking at the index or when querying ALL_IND_EXPRESSIONS on the table TABLE_NAME is the following:
  
  SQL Dev index is: NVL("CREATION_DATE",'01-JAN-4000')
  
  The Index of over SQL is: NVL ('CREATION_DATE2', TO_DATE ('4000-01-01 00:00:00 ',' yyyy-mm-dd hh24:mi:ss'))))
  
  Why are they not compatible?
  
  If I then fill up with 1 million rows and fill creation_date/creation_date2 with enough data to make no small enough value I then collect statistics on the table and explain the plans of selects on creation_date/creation_date2 via SQL dev. and SQL Plus what I see, it's different
  
  
  -Select
  NVL SELECT (creation_date, ' 01 - JAN - 4000 "")
  OF eagadev.paul_test
  where nvl (creation_date, ' 01-JAN-4000') = ' 01-JAN-4000';
  
  SQL Dev explain Plan
  Use PT_IDX1
  
  
  More SQL explain Plan:
  
  SQL > SELECT count (1)
  2 FROM eagadev.paul_test
  3 where nvl (creation_date, ' 01-JAN-4000') = ' 01-JAN-4000';
  
  COUNT (1)
  ----------
  100
  
  
  Execution plan
  ----------------------------------------------------------
  0 SELECT STATEMENT Optimizer = CHOOSE (cost = 30 card = 1 bytes = 8)
  1 TRI 0 (AGGREGATE)
  2 1 INDEX (LINE SCAN) OF "PT_IDX1" (NO ONE) (cost = 30 Ca
  RD = 5825 bytes = 46600)
  
  Statistics
  ----------------------------------------------------------
  14 recursive calls
  0 db block Gets
  8 compatible Gets
  0 physical reads
  Redo size 60
  224 bytes sent via SQL * Net to client
  276 bytes received via SQL * Net from client
  2 SQL * Net back and forth to and from the client
  0 sorts (memory)
  0 sorts (disk)
  1 rows processed
  
  
  -Select2
  
  SELECT count (1)
  OF eagadev.paul_test
  where nvl (creation_date2, ' 01-JAN-4000') = ' 01-JAN-4000';
  
  SQL Dev explain the Plan:
  FULL analysis
  
  More SQL explain Plan:
  
  SQL > SELECT count (1)
  2 FROM eagadev.paul_test
  3 where nvl (creation_date2, ' 01-JAN-4000') = ' 01-JAN-4000';
  
  COUNT (1)
  ----------
  100
  
  
  Execution plan
  ----------------------------------------------------------
  0 SELECT STATEMENT Optimizer = CHOOSE (cost = 28 card = 1 bytes = 8)
  1 TRI 0 (AGGREGATE)
  2 1 INDEX (LINE SCAN) OF "PT_IDX2" (NO ONE) (cost = 28 Ca
  RD = 5825 bytes = 46600)
  
  
  
  
  
  Statistics
  ----------------------------------------------------------
  14 recursive calls
  0 db block Gets
  8 compatible Gets
  0 physical reads
  Redo size 60
  224 bytes sent via SQL * Net to client
  276 bytes received via SQL * Net from client
  2 SQL * Net back and forth to and from the client
  0 sorts (memory)
  0 sorts (disk)
  1 rows processed
  
  
  The select on creation_date2 says plan is different between Dev SQL and SQl Plus. SQL Dev does not seem to be implicitly convert the '' 01-JAN-4000 in a date.
  
  Their behaviour are not the same?

  Then the problem may be due to different configurations of NLS between SQLPlus and SQLDeveloper, try to check the date format configured NLS in SQLPlus and SQLDeveloper.

  You can change the date format in SQLDeveloper, NLS of:

  Tools-> preferences - ZDatabase-> NLS

• How to view statistics

  Hello
  11 g R2 on Win 2003 with sqldeveloper 1.5.4
  I can see the plan to explain it, but no statistics.
  How do I view statistics when runing a query? Information such as:

  recursive calls
  db block gets
  consistent gets
  physical reads
  redo size
  bytes sent via SQL*Net to client
  bytes received via SQL*Net from client
  SQL*Net roundtrips to/from client
  sorts (memory)
  sorts (disk)
  rows processed

  Thank you.

  For those, you must run an Autotrace (F6), the icon next to the Plan to explain (F10).

  Have fun
  K.