11g RAC - RAW Devices Configuration

Similar Questions

• Can the sys admins save raw devices?

  Platform: Solaris 5.10
  DB versions: 11g
  
  As raw devices are not from the OS to the format sys admins can save raw devices? If they cannot, then, this means that we completely rely on backups RMAN to our DBs CARS. Right?

  resistanceIsFruitful wrote:
  Platform: Solaris 5.10
  DB versions: 11g

  As raw devices are not from the OS to the format sys admins can save raw devices?

  Yes. They can use the command dd (disk image) backup - where If (input file) is the raw device and of (output file) is a tape drive or a backup device.

  If they cannot, then, this means that we completely rely on backups RMAN to our DBs CARS. Right?

  This should be the preference. Backup of a raw device using dd does not guarantee a database or the cluster that can be restored - as other processes are / can be write data, while dd is playing it to the backup. It doesn't have the same management and control functions and the flexibility of rman.

  Do not use rman would be a mistake to IMO.

• Oracle 11g Release 2 assumption of raw devices to the CCR

  Hi I need one of my db migrate to 10.2.0.3 for 11g release 2
  now, I got to know that there are some problems with raw devices in oracle 11g R2 rac
  so I want the suggestions of experts
  Thanks in advance

  If you are upgrading a database in place (i.e. without reconstruction, without import-export), Oracle has no other choice than to support the application of raw device.
  However, when you install the Clusterware 11.2 it cannot be installed on raw devices - one is not supported!

  Oracle Support Document ID 578455.1 ' announcement of of-Support devices version 12 G GROSS ' announces support to 12 g.

  Hemant K Collette

• Configuration of the laptop for R12 with node 2 11g RAC database

  Hi all
  
  I plan to buy a new computer dell laptop XPS to install Oracle Apps R12 with node 2 11g RAC DB by creating virtual machines.
  
  Configuration:
  

  generation of Intel processor i5-2410 2 (2.3 GHz, core 2, 3 MB cache)
  8 GB RAM DDR3
  750 GB HARD drive

  this configuration will be sufficient for the highest installation? RAM? Processor?
  
  I've read a few threads over on this forum where people have recommended for installation of single application (11i or R12), this kind of configuration can be enough.
  
  Please answer.
  
  Thanks in advance.
  
  Brij

  I plan to buy a new computer dell laptop XPS to install Oracle Apps R12 with node 2 11g RAC DB by creating virtual machines.

  Configuration:

  generation of Intel processor i5-2410 2 (2.3 GHz, core 2, 3 MB cache)
  8 GB RAM DDR3
  HARD DRIVE 750 GB

  This configuration will be sufficient for the highest installation? RAM? Processor?

  You need at least 4 GB for installation on a single node R12. So, if you plan to have 3 VMs (1 application node) and 2 RAC nodes or 4 virtual machines (2 application and 2 RAC nodes), then I suggest that you add more memory. However, your RAM should work, but don't expect (if you go with 3 VMs) but do not expect the performance to be perfect.

  Thank you
  Hussein

• Raw devices not visible to the installer of the grid

  Version: 11G Rel 2
  Operating system: Solaris 5.10
  
  Number of nodes: 2
  
  We decided to use the oracle user and no network user as the owner of grid infrastructure because it is just a test installation.
  
  We have done the following on Node1

  # ln -s /dev/rdsk/c4defs0  ocr
  # ln -s /dev/rdsk/c4defs1  vote
  # ln -s /dev/rdsk/c4defs3  data1
  
  
  # chown root:oinstall /dev/rdsk/c4defs0
  # chown oracle:oinstall /dev/rdsk/c4defs1  -- But still owner remains root!
  # chown oracle:oinstall /dev/rdsk/c4defs3
  
  # chmod -R 777 /dev/rdsk/c4defs0
  # chmod -R 777 /dev/rdsk/c4defs1
  # chmod -R 777 /dev/rdsk/c4defs3
  
  
  # chmod -R 777 ocr
  # chmod -R 777 vote
  # chmod -R 777 data1

  For the storage attached to node2, we did

  # ln -s /dev/rdsk/c0defs1  data2
  # chmod -R 777 /dev/rdsk/c0defs1 -- for data files
  # chmod -R 777 data1

  During the drive for the installation of grid affecting the disc OCR and voting, part
  For

  Change Disk Discovery Path

  I entered

  /dev/rdsk/

  But it is to recognize only the voting devices and data1. OCR (/ dev/rdsk/c4defs0) cannot be seen. No idea why?
  
  Another thing, I noticed. I don't see node2 (dev/rdsk/c0defs1) disk device. Shouldn't this be included as well at this point?

  Hi Steve,.

  RAW or block devices are supported only when upgrading an existing installation using the already configured partitions. On new installations, using raw or block device partitions is not supported by Oracle Automatic Storage Management Configuration Wizard (ASMCA) or Oracle Universal Installer (YES), but is supported by the software if perform you a manual configuration.

  The above information is corret.
  In my opinion, which is always supported in case of upgrade.
  I did not write this because being support block and raw devices on this version.
  With the release of Oracle 11 g 2 (11.2) database and RAC Oracle 11g release 2 (11.2), using the database Configuration Wizard or Setup to store files directly on the block or raw devices Clusterware Oracle or Oracle database is not supported.

  >

  And by manual configuration, they hear

  # chown grid:asmadmin /dev/rdsk/cxtydzs6
  # chmod 660 /dev/rdsk/cxtydzs6
  

  Right?

  Since the oracle user is the owner of the software grid in my case it would be

  # chown oracle:oinstall /dev/rdsk/cxtydzs6
  # chmod 660 /dev/rdsk/cxtydzs6
  

  Yes...
  When the grid is that you are the owner to install Oracle Grid Infrastructure, in case a manual installation using raw for ocr and vote oinstall is the Group

  Kind regards
  Levi Pereira

• raw devices problem

  Hello, I did a test Oracle RAC on Vmware ESX model, so it is two machines of Centos 5.10 with shared disks. I'm using raw devices for ocr, voting records and data, and without the ASMLib library.

  After you install the clusterware and Oracle 10 g (2), everything was OK so far, when after the reboot of a node, not more shared disks exist anymore. For each related raw disk comand I got an error:

  #raw - qa

  Could not open the raw device ' / dev/rawctl "(no such file or directory)

  #service rawdevices restart

  Allocation of raw devices:

  / dev/RAW/raw1->/dev/sdb1

  Could not open the raw device ' / dev/rawctl "(no such file or directory)

  Have not found anything related to this in/var/log Pourrait be happy if someone could explain how to solve this problem.

  Yes, but keep in mind that the tree view of the devices is not only rebuild after a reboot of the system, but also the order of devices may be different, so by example/dev/sdc1 may not be the same after a system restart if you fix or remove another device or plug any USB. That is why UDEV must be configured to use the unique device ID to map storage devices requiring a persistent device naming.

  For example, you can use the following entry to map for example current/dev/sdc1 on/dev/vote1

  KERNEL == "sd *", BUS is "scsi", PROGRAM is "/ sbin/scsi_id" RESULT is "361a98001686f6959614a453133524171", NAME = "vote1", OWNER = 'oracle', GROUP is 'oinstall", MODE ="0640.

  It is no longer necessary, not for Oracle, which use the O_DIRECT flag when opening a device, and no matter if it is a device or partition to use raw devices. But if you really want to match names to raw device, you can use the following text:

  ACTION == "add" KERNEL == "sd *", PROGRAM == "/ sbin/scsi_id", RESULT == "361a98001686f6959614a453133524171", RUN += "/ bin/raw/dev/raw/raw1 %N.

• Deployment of Oracle 11g RAC Release 2 with IBM midsize V7000

  At and deployment 11g rac with IBM midsize V7000 san with aix 6.1
  Is this possible using Oracle CRS without going to the OS like GPFS cluster

  lbn76 wrote:
  At and deployment 11g rac with IBM midsize V7000 san with aix 6.1
  Is this possible using Oracle CRS without going to the OS like GPFS cluster

  Oracle Grid these days use ASM starts for vote and register files storage. So no need for a cooked cluster file system that must be mounted by all nodes. No need either directly with a gross storage logical unit number.

  ASM uses the LUN of the matrix of storage/SAN directly (as devices of rough blocks).

  Generally however, there are several I/O paths to that LUN on the SAN - in which case there is more than 1 device level o/s for this LUN. Multipath driver s/w addresses this and creates a logical (and redundant) device that sits at the top of several devices of i/o path to the level of the o/s.

  You configure ASM to use devices of paths multiples (as devices of rough blocks) - and applying exterrnal, normal or high redundancy necessary when creating starts on these logical multipath devices.

  The grid s/w and RAC database instances can then use ASM starts as cluster storage.

  This makes storage much easier layer management - and also provide a number of useful features too.

  This is also the approach recommended by Oracle Corp. for RAC - use ASM under the layer of management of storage between storage LUNS in the table and one side and CARS on the other.

• Output device - configuration of the handle bar not to change

  This question has me left speechless.

  Usually, I run the application using device-Release configuration. Yesterday, after construction with Simulator-debug, I turned back on build to release device configuration.

  Now I find that whenever I build the application, the active path in the bar - descriptor.xml automatically converted to the x 86 instead of arm/o.le-v7/. I can't run the application in release device mode now because of this automatic conversion.

  
         armle-v7
         lib/libBBCommonDatabase.so
         TestApp
     

  Clean up the project and build the project in the peripheral output mode before the race.

  I also facing the problem when I did as out of fashion Simulator to the peripheral debug mode.

  So try this one maybe it will help you.

• How to find the logic of a RAW device unit ID-number

  Hello

  I am trying to create a workflow that will give me information about a virtual machine:

  • RAM
    
  • CENTRAL PROCESSING UNIT
    
  • Disk configuration
    
    • Size, SCSI Bus
      
    • ..
      

  My problem is to understand the LUN ID to a RAW device in a virtual computer.

  I think, a RAW device is listed as the VcVirtualDisk type too?

  Anyone know how to determine the LUN ID of a RAW device?

  Thank you

  Marc

  Hello!

  This should help you:

  http://communities.VMware.com/message/2012069

  See you soon,.

  Joerg

• Boot partition of windows 7 64 bit as RAW device in VMware workstation 7.1

  My setup is a laptop to dualboot Lenovo W510 with Windows 7 64 bit on a 200Gig partition and Debian GNU/Linux Squeeze on partition of 100Gig. I've got workstation 7.1 installed on Linux and it works very well for virtual machines. The hard drive is a SATA.

  I'm trying to start the 64-bit Windows 7 as RAW device partition. After different assignment on this forum as well as on the web, I had gathered that I needed to change the entries from scsi to ide in the .vmx and too hard. Even with these changes, I get a BSOD 7 b error when starting windows VM. Previously, I had been able to do it on a Lenovo T400, with exactly the same Setup to dual-boot.

  I also read a KB that SATA is not supported as a CRUDE device. Linux guests see the SATA as SCSI device.

  Anyone has the fresh ideas on this or have met and solved this problem?

  I stick it on my site: http://sanbarrow.com/p2v2008r2.html#patching

  It shows how to configure the vmx file and how to fix the 7B...

  Patching - easy Variant

  This procedure is used when otherwise the VM would start in a kind of bluescreen 7B.

  This means that VMware is the target and that you configure the computer virtual with VirtualHardware type 7.

  It applies only if the virtual disk is configured like this

  scsi0.virtualDev = "lsisas1068".

  scsi0:0. FileName = "imported.vmdk".

  On the platform target start the virtual machine, newly created in the Windows 7 recovery CD.

  After the initialization is complete, click CMD and run regedit.

  Select HKLM, then Menu > file > load hive.

  Navigate to the Windows directory on the virtualdisk > system32\config and select the file SYSTEM.

  When asked for a name enter p2v.

  Then find the key

  HKLM\p2v\ControlSet001\services\LSI_SAS

  and the startup type of the REG_DWORD value 0

  Select HKLM\p2v, and then click Menu > file > unload hive.

  You can now close regedit and restart.

  The next time you start, you can start in the 2008 R2 or Win7 system. He should find the drivers and probably need to reboot when finished.

  _________________________

  VMX-settings- WS FAQ -[MOAcd | http://sanbarrow.com/moa241.html]- VMDK-manual

• Best practices - virtual disk of Raw device mapping vs

  Hello

  A small question about the configuration. What would be described as the best practice to configure a comment server that required a large amount of disk space? What is the preferred option:

  1 - create a virtual machine with a mapping of raw device to the LUN (on a san)

  or

  2 - create a virtual machine with another large virtual disk that is in a data store on the LUN (on a san)

  Thank you

  Dave

  Kader you are limited with 2 TB in any case (except using the iSCSI software from within the guest operating system), the only arguments are:

  (1) use RDM if it is possible that you will connect to this physical machine LUN. Or you already have the unit number logic of a physical machine as a large file server.

  (2) VMFS use in all other situations - VMDK give flexibility, you can migrate the virtual machine where you want.

  ---

  MCSA, MCTS, VCP, VMware vExpert 2009

  http://blog.vadmin.ru

• HP EVA4400 + ESX 3.5.0 + discs Windows 2003 Raw Device Mapping

  Hi all

  I'm testing a configuration to help:

  • HP EVA 4400

  • ESX 3.5

  • Blade HP c-class

  Verification of compatibility EVA4400 on-site HP Spock, I found this note in the section Notes General Guest OS (1): "(7) Raw Device Mapping is not supported on Windows 2003 Guest Operating System load."

  I would use snapshot EVA (BC) to a Windows 2003 Server (not R2) for backup.

  In the VMware HCL, I have not found confirmation of this note.

  1. does anyone know if the note concerned only W2K3 or same W2K3 R2?

  2. "is there anybody out there" who uses this configuration? EVA4400 is: + W2K3 in ESX 3.5 + RDM + instant. Any problem?

  Thank you very much

  Max

  The restriction was lifted (don't know why he never existed - perhaps no trial):

  Release notes for HP StorageWorks 4400 Enterprise Virtual Array (XCS 09006000)

  Reference number: 5697-7997, second edition: January 2009

  * Configuration of VMware for the raw device mapping details

  RAW Device Mapping only is not supported with Windows 2003 Machine virtual guest operating system and versions prior to version ESX 3.5.0.

• How to copy controlfile placed in RAW device

  How to copy the raw device control.
  We have 9i rac. All control files are placed in the device name/dev/control1 rough, / dev/control2.
  
  My question is if/dev/control1 is damaged by any reason, then how to copy/dev/control2. for/dev/control1.
  Normal os copy command will work?

  You can use the dd on UNIX OS command to copy the raw device on another raw device files.

  Refer to this: http://download.oracle.com/docs/cd/B10500_01/server.920/a96572/osbackups.htm

• Solaris RAW device

  Hello
  
  I am new to solaris and I want to create a raw device, how can I do?
  
  Thank you

  Hello

  To configure raw Partitions or raw logical Volumes on Solaris environment

  See this link
  http://download.Oracle.com/docs/HTML/B10766_08/presolaris.htm#BABJCHDJ :)

• How to check the space on the asm and raw device?

  Hi all
  
  I want to know how to check the space on the device of the DSO.
  We use a feature of the asm, which is on a raw partition.
  Our archive is full. I need to check the space on asm archve partiotion.
  How can I check the destination space archive?
  We are using partition raw device for crs and voting system.
  How can we check the space on the raw device and how do know us which device is currently in use and are not in service
  
  Thanks in advance

  4 - 5 raw partitions...
  What version of oracle?
  Install the OCR + VOTE in gross unit better than OCFS2. If you use 10g to 11 GR 1 material.

  If use 11 GR 2, take the OCR + VOTE in ASM disk group.

  You can check what raw OCR device:
  $ ocrcheck

  VOTEDISK:
  $ crsctl query css votedisk

  On 10g if using the raw device for OCR + votedisk: recommend 1 file = 250M
  11g recommend 1 file = 512M

  Good luck