Triple mirrored in normal redundancy

Similar Questions

• Normal redundancy and physical disks

  Version of the grid: 11.2.0.3
  Platform: Oracle Enterprise Linux 6
  
  
  I'm a bit confused about the term "Two-way mirrored" in the documentation of the DSO.
  
  
  If I want to keep 1 tera bytes of business data in an ASM diskgroup with NORMAL redundancy.

  Should I need 2 physical Disks of 1TB each ?
  or 
  Should I need 3 physical Disks of 1TB each ?

  ASM is not a software RAID. ASM works based on file extensions and provides data redundancy by using the free space on drive failure of disk groups. By default, each disc is a group of failure. For a mirror (normal) 2 channels, you need at least 2 rupture discs groups; for the 3 - way (top), you need at least 3. You can also create a 2-way mirror with 3 drives, in which case data will be written alternating. Having ASM normal redundancy in your example you must have at least two drives of 1 TB, but you can also use 3 and will have more space. You can also use an external RAID and use no redundancy ASM (external), although ASM will give you better performance and flexibility because you can add disks to increase the space without having to rebuild your storage space.

• Is this normal in "normal redundancy?

  The GI version: 11.2.0.4 (RAC 3-node)

  OS: Oracle Linux 6.4

  Storage EMC Symmetrix VMAX

  I created a diskgroup with normal redundancy as shown below.

  10 each = 5000GB 500 GB LUN

  CREATE DISKGROUP DATA_DG1

  NORMAL REDUNDANCY

  DISC

  "ORCL:DATA_D1,"

  "ORCL:DATA_D2,"

  "ORCL:DATA_D3,"

  "ORCL:DATA_D4,"

  "ORCL:DATA_D5,"

  "ORCL:DATA_D6,"

  "ORCL:DATA_D7,"

  "ORCL:DATA_D8,"

  "ORCL:DATA_D9,"

  'ORCL:DATA_D10 '.

  ATTRIBUTE 'au_size' = "4 M";

  DiskGroup created.

  Given that it was normal, I was expecting half of gross space 5000 GB = 2.5 terabytes of usable space. But, the query below shows the output, I myself only 2.19 terabytes of usable (2249,86 GB) space. About 250GB short of what I expected. So where's my advantage 250 GB?

  SQL > select name, TYPE, 1024/1024/TOTAL_MB, FREE_MB/1024/1024, USABLE_FILE_MB/1024/1024 of v$ asm_diskgroup where name = 'DATA_DG1 ';

  NAME TYPE TOTAL_MB/1024/1024 FREE_MB/1024/1024 USABLE_FILE_MB/1024/1024

  ------------------------------ ------ ------------------ ----------------- ------------------------

  NORMAL DATA_DG1 4.8828125 4.8825531 2.19713593

  The amount of usable file space is equal to the amount of free disk space subtracted from the amount of disk space required to restore redundancy after failure divided by the level of redundancy of the disk group that is in your case.

  Run:

  select trunc((free_mb - required_mirror_free_mb) / 2) as useable from v$asm_diskgroup where name='DATA_DG1';
  

  Source: http://prutser.wordpress.com/2013/01/03/demystifying-asm-required_mirror_free_mb-and-usable_file_mb/

• 11 GR 2 Grid Infrastructure install - with normal redundancy ASM disk group requires 3 disks?

  
  Hello

  I want to install 11 GR 2 grid Infrastructure in a new environment in cluster - 2 nodes). I use the 11.2.0.4 GI Installer on Oracle Linux 6.5 x 86-64

  However when it is in the installer of GI, it comes to choose the storage Option - I choose ASM disk groups.

  Then change the path of discovery to pick up my candidate ASM disks.

  I enter the disk group name SIR give Normal redundancy, choose 2 discs of candidate and then press < next > to Setup.

  However, I gives me the message

  [INS-30510] Insufficient number of ASM disks selected

  The number of selected disks was inadequate for the level of redundancy selected

  For a disk group with redundancy level "Normal" disks at least "3" are recommended

  I always thought that Redunancy Normal required only 2 discs and high redundancy requires 3 disks?

  No idea what is gpoing?

  Thank you

  Jim

  Consult the following notes for how to move:

  Note 1638177.1 and 428681.1

  Concerning

  Thomas

• Half Rack exadata with DBFS_DG have normal redundancy?

  My understanding is that one install scripts command automatically the value of the level of redundancy for the diskgroup DBFS_DG. I think I had actually foud this in the code of gridsteps.pm function CreateASMDiskgroups() onecommand. For an eighth and a quarter rack, it's Normal for half and full, it is high. This is to ensure that critical OCR and vote the files are backed up as far as possible. But recently I came across several facilities half rack with DBFS_DG redundancy as usual. Is it possible to replace this or does this mean that the grid software has been reinstalled manually at some point?

  It's about x 2-2 half baskets. Grateful if someone knows if there is a way, it could happen through an implementation of the standard onecommand.

  OneCommand creates DBFS_DG in normal redundancy each time.  If you are on a half or a full rack with high diskgroup redundancy, he moves the OCR and files to this vote diskgroup.  This happens at the stage of CreateASMDiskgroups of onecommand.

• Redundancy of the Normal ASM - three stores, three groups of failure

  Hello

  I have a question about groups of failure in ASM. I have an ASM diskgroup with normal redundancy and records of three different preparations. Each storage disks are in separate groups of failure (see screen below). I know what ASM allocates to a point as it allocates a copy of copy and the primary mirror and mirror copy is placed on a different disk that is part of a group of different failure. And it is understood for the configuration with two storage (or controllers) and two different groups of failure. But what with configuration where are three storages and failure groups? My question is: If I failed three groups if new extensions are allocated are mirrored are placed on the other two discs are it is a part of a group of different failure? Or maybe this kind of configuration (with three groups of normal redundancy ASM failure) is not supported?

  

  I know what ASM allocates to a point as it allocates a copy of copy and the primary mirror and mirror copy is placed on a different disk that is part of a group of different failure. And it is understood for the configuration with two storage (or controllers) and two different groups of failure. But what with configuration where are three storages and failure groups? My question is: if I have three groups of failure, if new extensions are allocated are mirrored are placed on the other two discs are it is a part of a group of different failure? Or maybe this kind of configuration (with three groups of normal redundancy ASM failure) is not supported?

  No matter how much you failgroup. Normal redundancy requires at least 2 Failgroup.

  Oracle will allocate primary measure in a failgroup and place the mirror of the scopes in an another failgroup.

  The order does not matter since you always have a copy in a different failgroup.

  See example below (d = data extended)

  FG1 |   FG2 | FG3

  ========|===============|=========
  

  d1      |   d1_copy:

  |   d2          | d2_copy

  d3_copy |               | d3

  When you create a Diskgroup without specifying the failgroup, Oracle will automatically create a failgroup to each diskgroup. So, if you create a normal with normal redundancy Diskgroup and ASMDISK 10, ASM will create 10 failgroups.

  You have 3 warehouses and 3 Failgroup

  If you loss storage (h/f) ASM will tolerate because you measure mirror in any other failgroup (h, l, square).

  If you loss both storage (h/f) or two different storage disk/controller at the same time ASM will not tolerate. (Need to restore a backup database)

  Your point of failure: you must make sure that the two storages/controllers/disks will still be available.

  If your concern is: can I have a situation where two storage can do fail (or cut) at the same time. Then, you need high redundancy, because this configuration will tolerate you two storages loss (h/f) at the same time and Diskgroup will remain active.

• QUESTION OF REDUNDANCY NORMAL ASM

  Hi all
  
  Document Oracle ASM NORMAL QUESTION of REDUNDANCY [1480954.1 ID] says the statement, I find it difficult to understand the line which is marked in italics below.
  
  To determine the appropriate size of a diskgroup which is protected with redundancy of ASM, sufficient free space must exist in the diskgroup

  so that when a drive goes down, it can automatically rebuild the contents of the drive failed on another drive in the diskgroup. The database should stay online if sufficient space is available

  . The amount of space required to ensure ASM can restore redundancy if there is no disc is in the REQUIRED_MIRROR_FREE_MB column in the view V$ ASM_DISKGROUP. The amount of free space that can be used safely in a disk group, taking into account, mirroring and still be able to restore redundancy after a drive failure is in the USABLE_FILE_MB column in the view V$ ASM_DISKGROUP. USABLE_FILE_MB must always be greater than zero. If USABLE_FILE_MB falls below zero, most records should be added to the disk group.
  
  DOUBT: why what he have reconstitute the content of the disc failed on another drive in the diskgroup when scopes in mirror and I do not understand when the disc itself is failed how they can access it and rebuild its contents to another drive.
  
  
  Concerning

  DOUBT: why need restore the contents of the drive failed on another drive in the diskgroup when there mirrored scopes

  because the data on the failed drive is not more in the opposite direction (no additional copies) and to ensure a level of normal redundancy, we'll build a new mirror copy of data

  I do not understand when the disc itself broke down how they can access it and rebuild its contents to another drive.

  data always available, mirrored

  to explain it simply:

  everything is stored with 2 copies (original + mirror)
  When 1 is lost, a new copy will be rebuilt among others the 1, so there will be 2 copies again

• How OCR mirroring works

  Hi gurus,

  According to the Oracle docs "disk groups normal redundancy provide 3 disk files with right to vote, 1 OCR and 2 copies (a primary and a secondary mirror).»

  Does this mean there are 2 OCR in the diskgroup?. Then,

  Where the OCR Elementary and mirror created in this case? Are they were distinct failure group? So how can I check the same through asmcmd?

  This is my environment set in place

  [grid@xxxx01 ~] $ ocrcheck

  Status of the Oracle Cluster registry is:

  Version: 3

  Total space (in kilobytes): 262120

  Used space (in kilobytes): 2800

  Amount of available space (KB): 259320

  ID: 1074808153

  Device file name / : + OCR_VOTE

  Verifying the integrity of file/device succeeded

  File/device not configured

  File/device not configured

  File/device not configured

  File/device not configured

  Checking cluster registry integrity succeeded

  Logical corruption control bypassed due to the non-privilegie user

  My + OCR_VOTE is in normal redundancy, while watching the level to ASM

  [grid@xxxx01 ~] $ asmcmd

  ASMCMD >

  ASMCMD > cd + OCR_VOTE

  ASMCMD > ls

  ehrptpi-cluster.

  ASMCMD > cd xxxxxi-cluster.

  ASMCMD > ls

  ASMPARAMETERFILE /.

  OCRFILE /.

  ASMCMD > cd OCRFILE /

  ASMCMD > ls-l

  Name of type Redund time scratched Sys

  OCRFILE MIRROR RUDE JAN 07 20:00 Y REGISTRY.255.763526561

  ASMCMD >

  Here I can see OCR only one file. So what is meant by 1 OCR and 2 copies?

  Can someone shed some light on how OCR mirrored actually works?
  

  Thank you and best regards,

  Mahi

  Hi Manon,.

  you have different KFFXP because there are two different files. 253 is you spfile asm and 255 is your ocrfile. Asmcmd pass in your diskgroup to ocr_vote to see the two files. It should look like this:

  ASMCMD [+ SYSTEMDG/rac-entw] > ls

  ASMPARAMETERFILE /.

  OCRFILE /.

  ASMCMD [+ SYSTEMDG/rac-entw] > ls ASMPARAMETERFILE.

  REGISTRY.253.804243265

  ASMCMD [+ SYSTEMDG/rac-entw] > ls OCRFILE.

  REGISTRY.255.804243269

  ASMCMD [+ SYSTEMDG/rac-entw] >

  The recommendation for an another OCR second or even third is more fault tolerant. With an OCR (mirrored with two copies), you can lose a copy but not all diskgroup. If you have more on other OCR starts you might survive even the loss of a diskgroup together.

  However, since your votefiles are stored on the disks in the Group ocr_vote on your cluster would break even if you lose this group. However, you have an up-to-date version of your ocr, compared to a version of several hours of your backup of the ocr.

  Beside that, I see personally very little benefit in addition to a mirror OCR.

  What do you mean with master OCR?

  Concerning

  Thomas

• ASM level of redundancy

  Dear all,
  at the level of normal redundancy, I have 2 asmdisks in a diskgroup (not create failgroup), each disk size is: 5 G, guess I have 4 GB of data, then how the data are published in the diskgroup?
  
  Thank you

  TyL wrote:
  Dear all,
  at the level of normal redundancy, I have 2 asmdisks in a diskgroup (not create failgroup), each disk size is: 5 G, guess I have 4 GB of data, then how the data are published in the diskgroup?

  You don't need to explicitly create a group fail. Each disk will be attributed to his own failure group by default and therefore, the data would be mirrored on two drives by oracle automatically.

  Aman...

• ASM and mirroring

  DB version: 11.2
  Version of the operating system: Solaris 5.10
  
  System administrators have provided us with a disk array with a total capacity of 700 GB for shared storage. I created two slices of 1 GB for disc OCR and VOTE.
  
  Now, I find myself with 698 GB I would use to create a group of disks for data files. Using

  format 

  command I can create several tranches from there or I could simply create a slice of 698 GB. Should which option I choose?
  
  For normal redundancy in ASMCA option, you must have at least 3 slices.
  
  ASM actually offers mirroring? Lets say that one of the slices we used to create the disk group had a hardware failure. Will we ever know of this failure?

  Hello

  System administrators have provided us with a disk array with a total capacity of 700 GB for shared storage. I created two slices of 1 GB for disc OCR and VOTE.

  You use the storage?

  If so, I sugest you use conventional RAID on the storage systems, if you have the choice.

  Create LUNS/disks for the first diskgroup to store VOTING and OCR.

  Re: PROT-30: Cluster Oracle location register to add is not accessible

  command I can create several tranches from there or I could simply create a slice of 698 GB. Should which option I choose?

  I recommend create you discs (slice) or ONE of the 100GBytes. This will help you in the future management of the environment.
  If you use more than one of THEM, this will make a big difference.

  For normal redundancy in ASMCA option, you must have at least 3 slices.
  ASM actually offers mirroring? Lets say that one of the slices we used to create the disk group had a hardware failure. Will we ever know of this failure?

  Yes ASM provide mirroring, but if you use a disk (lun) not supposed to create mirroring in the same drive. You will need at least two disks in different tables.

  Kind regards
  Levi Pereira

• Redundancy of the ASM

  Hello
  
  I have a 1 TB raid 10 storage san for RAC DB installation and wanted to ask you what redundancy is preferred for my installation. External/Normal/high?
  And as I know if I want external I don't need to partition my 1 TB storage, if I put it as Normal I have to partition to 2 partitions and if I put at the top, I have to partition the storage of 3 parts, am I right?
  
  Thank you

  I'd go for the normal redundancy, it's a good compromise. Make sure you have 2 independent physical devices. Partitioning is not clear to me, what do you mean? It is not supposed to partition a physical device and set the redundancy of the DSO on that, if you lose this device the mirror (s) have also disappeared.
  External redundancy can be used if you are mirroring a typical RAID implementation. I wouldn't mix redundancy RAID and ASM, choose one of them.

  Werner

• WITH REDUNDANCY ASM DISKGROUP

  Hi all

  
  

  REDUNDANCY NORMAL == NEED 2 GROUPS of FAILURE (I'm clear)
  

  
  

  I am trying to create diskgroup with REDUNDANCY NORMAL (only with 2 raw devices).
  

  Command line == > I can't create disk with REDUNDANCY NORMAL group

  Through DBCA == > I can.
  

  
  

  select GROUP_NUMBER, name, ALLOCATION_UNIT_SIZE AU_SZ, STATE, TYPE, TOTAL_MB, FREE_MB, OFFLINE_DISKS from v$asm_diskgroup;
  
  GROUP_NUMBER NAME                                AU_SZ STATE       TYPE     TOTAL_MB    FREE_MB OFFLINE_DISKS
  ------------ ------------------------------ ---------- ----------- ------ ---------- ---------- -------------
             1 ASM_DG_DATA                       1048576 MOUNTED     NORMAL      20472      14142             0
             2 ASM_DG_FRA                        1048576 MOUNTED     EXTERN      10236       7143             0
  
  

  
  SQL> create diskgroup DATA NORMAL REDUNDANCY
    2  failgroup disk1 Disk
    3  '/dev/raw/raw5' name data1,
    4  '/dev/raw/raw6' name data2;
  create diskgroup DATA NORMAL REDUNDANCY
  *
  ERROR at line 1:
  ORA-15018: diskgroup cannot be created
  ORA-15072: command requires at least 2 failure groups, discovered only 1
  
  

  But through dbca, I can create diskgroup only with 2 drives

  

  SQL> select GROUP_NUMBER, name, ALLOCATION_UNIT_SIZE AU_SZ, STATE,  TYPE, TOTAL_MB, FREE_MB, OFFLINE_DISKS from v$asm_diskgroup;
  
  
  GROUP_NUMBER NAME                                AU_SZ STATE       TYPE     TOTAL_MB    FREE_MB OFFLINE_DISKS
  ------------ ------------------------------ ---------- ----------- ------ ---------- ---------- -------------
             1 ASM_DG_DATA                       1048576 MOUNTED     NORMAL      20472      14142             0
             2 ASM_DG_FRA                        1048576 MOUNTED     EXTERN      10236       7143             0
             3 ASM_DATA                          1048576 MOUNTED     NORMAL      40946      40844             0
  
  

  SQL> select a.disk_number, b.group_number, a.path, b.name, b.type from v$asm_disk a, v$asm_diskgroup b 
  where a.group_number=b.group_number;
  
  DISK_NUMBER GROUP_NUMBER PATH                   NAME                           TYPE
  ----------- ------------ ---------------------- ------------------------------ ------
            0            3 /dev/raw/raw6          ASM_DATA                       NORMAL
            1            3 /dev/raw/raw5          ASM_DATA                       NORMAL
            0            2 /dev/raw/raw3          ASM_DG_FRA                     EXTERN
            1            1 /dev/raw/raw2          ASM_DG_DATA                    NORMAL
            0            1 /dev/raw/raw1          ASM_DG_DATA                    NORMAL
  

  ...

  ..

  My question/confusion is only 2 discs,

  I can create a diskgroup (redundancy = normal) using DBCA, but similar operation cannot perform through command line.

  DB version: 10.2.0.5 on RHEL

  Shows your screenshot of the ASMCA you assign disk/dev/raw/raw5 in FailGroup FG1 and disk/dev/raw/raw6 in FailGroup FG2 - two * different * FailGroups. Each disc is in a separate FailGroup.

  Your Interactive SQL command specifies only 1 FailGroup when there is to specify two FailGroups.

  Hemant K Collette

• How to avoid the brain while using external redundancy.

  Hi all

  When Oracle creates disk to vote who is in external redundancy he writes only in a single drive even if we provide 3 disks for the disk group. In this scenario, what will happen during the expulsion of node. Is there a scenario of split brain.

  Ajay

  Ajay Vijayan wrote:

  I would like to know more about it. How do the multiple vote available in external redundancy disk?

  You can not. If you want three voice files, use a group configured for normal redundancy.

• change of vote disk redundancy

  Hello Experts,

  During the installation of 12 c (12.1.0.2) RAC on AIX 7.3, we chose external redundancy for voting disk (with RAID 6 it works very well), but according to the guidelines of the Oracle, we would like to change to normal redundancy in which we will maintain three copies of raw disks to vote on three drives/devices.

  Please help us with MOS document or the directive to achieve the same. Thanks in advance.

  Abhijit-

  Create new diskgroup (CRS) with redundancy for the OCR and voting files. Check diskgroup mounted in all cluster nodes

  move OCR and file Vote again diskgroup

  Move the OCR, voting file, SPILE ASM Diskgroup (Doc ID 1638177.1) again

• ASM high redundancy on the quarter grid should be possible

  I was reading the white paper "Best practices for Database Consolidation on Exadata Database Machine" http://www.oracle.com/technetwork/database/features/availability/exadata-consolidation-522500.pdf and came across this statement
  
  On page 5
  
  «There is also the possibility of deploying a smaller pool consisting of a support in 2-2 quarter Exadata X for entry level of vital systems consolidation No.» This option is not recommended for the basic needs of the company because it does not fully support ASM disk groups high redundancy (full support for high redundancy ASM disk groups requires a minimum of Exadata X 2 - 2 half rack). »
  
  A quarter rack has 3 storage cells seems to me that a high redundancy should be possible. can someone shed some light on the origin and significance of this statement?
  Thank you
  Robin

  The reason that you can't high redundancy on all starts in a quarter of a rack is due to requirements for OCR/voting records when they reside within the ASM. For OCR and voting records be placed in high ASM diskgroup redundancy, it takes 5 failgroups. Because a quarter rack has only 3 cells, one of your start (probably DBFS_DG) must be normal redundancy to meet the OCR and voting disk conditions.

  It is possible to have DATA and RECO be high redundancy on the grid of a quarter, but they cannot hold OCR and voting disks.