Bare Metal Provisioning - physically needed machines

Hello together,
our company plans to use Bare Metal Provisioning soon.
We're thinking of the installation of the needed servers at the moment. According to the best practice guide, one needs a boot server, a stage server, a rpm repository and a software library.
In the document it's mentioned that oracle suggests to keep boot and stage server on one physical machine.
Can we also place the rpm repository or the software library (or both?) on this machine? Do we have any disadvantages when doing it this way? As servers are nothing cheap, we could save some money. But of course not if it would cause significant problems...
What do we have to consider about it?
Do you have any experience in consolidating those servers?
Thanks in advance,
Robert

Hi Robert,
If you have small number of targets, that would be find OMS and Reporsitory on same box. But for in future considering growth you may require distributed architecture.

Similar Messages

  • Information for All: Updated bare Metal provisioning best practices paper

    Hi All,
    The Bare Metal provisioning best practices paper has been updated and is now available on OTN at the following location:
    http://www.oracle.com/technology/products/oem/pdf/bmp_best_practice.pdf
    Thanks,
    Rajat

    Very nice, thank you.

  • Linux Bare Metal Provisioning with OEM Pack for Linux for RAC 11gGR2 OEL 5X

    Hi,
    Has anyone ever used the OEM Pack for Linux to provision Bare Metal Linux servers?
    See: http://www.oracle.com/technology/products/oem/omp_linux.html
    We are thinking of using this tool to provision our Linux Servers, because we are UL-Customers with OEL, and this Tool is free of Charge for UL- Premier and Basic Limited Customers.
    I have only found Documentation which seem to suggest that the OEM Linux Pack is supported with OEM Grid Control 10.2.0.4 and Oracle 10G ,
    but i am not sure if this Pack is supported for use with OEM Grid Control 10.2.0.5 and Oracle RAC 11gR2.
    I would appreciate your Help on this.
    Regards,
    Pascal

    user11999910 wrote:
    but i am not sure if this Pack is supported for use with OEM Grid Control 10.2.0.5 and Oracle RAC 11gR2.Yes, it is. The Linux Management Pack is built-into Grid Control and ships with 10.2.0.5. I believe you need to add a few patchsets to get full provisioning support for Database 11gR2. There is a My Oracle Support article that lists all the recommended patches for Grid Control.

  • SCVMM Bare Metal Deployment with SAN Boot

    Can VMM assign "boot targets" to hosts while deploying them? Or do I have to do it manually before initiating VMM deployment? Specifically talking about PowerEdge M420. But also interested if it is possible in general?
    The data above this text is pseudorandom, brace yourselves.

    Hi Sir,
    http://support.microsoft.com/kb/305547
    This article shows the key point is we need to make the LUN works properly like a local disk , it also means that we need to prepare HBA driver within WinPE image before the deployment .
    I have also read these blog regarding adding drivers for SCVMM Bare Metal Deployment:
    http://www.thecloudbuilderblog.com/blog/2012/1/26/scvmm-2012-bare-metal-provisioning-adding-drivers-to-winpe.html
    http://www.virtualizationadmin.com/articles-tutorials/microsoft-hyper-v-articles/installation-and-deployment/performing-bare-metal-installation-hyper-v-using-system-center-virtual-machine-manager-2012.html
    Best Regards,
    Elton Ji
    Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact [email protected] .

  • UCS B-Series bare metal OS FC Issues with failover

    My customer is seeing loss of access to FC attached SAN on b-series blades that have windows and Red Hat installed bare metal, VMware are fine just these 2 OS's.  It happens when the 6100 fails over to the subordinate.  They say they reboot the servers 4-5 times before they are live on the fabric again as intended.  They have many UCS systems in many environments with differing versions of firmware and too many variables to list here, looking to see if anyone seen and rectified this, possible drivers for 2008r2 and RH ??? 
    thx
    dave

    Dave,
    Proper fabric failover really pertains to the OS and how is deals with the storage platform paths. When UCS failsover the the fabric an MPIO driver on the OS must be tested and functional to achieve proper a connection status. When installing Windows or Linux and bare metal you will need to ensure that the correct MPIO driver is installed on the OS. This driver should be obtained from the storage admin in your enviroment and must be installed and tested for proper path failover prior ro any UCS failover testing. In the case of VMware, vSphere uses a native MPIO driver and works out of the box, Windows and Linux will require a MPIO driver to match this funtionalilty.
    I hope this helps,
    David

  • Is it okay to do a bare metal Linux Mint 17 install, or do I need to keep Windows 8 on my SSD drive?

    I will be installing a new Samsung EVO SSD drive in a new T540p.  I really don't use Windows anymore, so I'm wondering if I have to keep Windows 8 on this machine and set up a dual boot with the new Mint 17 on the new SSD - or - can I just install the Mint on the new bare metal SSD?  Can I do this and still have UEFI too?  I'd prefer that.  It would just be Mint 17 on the UEFI SSD.
    Or is this a crazy idea?   My old computer is dual boot and I just don't use the Windows OS anymore.  It's just uselessly using up my drive space.
    Solved!
    Go to Solution.

    The T540p is a really nice machine. I have Windows 8.1 running on mine. The display is fantastic. I'm sure you will really enjoy yours.

  • Bare-metal deployment of non-hyper-v host possible?

    Is it possible to use VMM 2012 R2 to perform a bare-metal deployment of a physical server which will not be a Hyper-V server? Just trying to find out if there's an opportunity to consolidate our build process to replace OSD or using the MDT. In this scenario
    I'm willing to consider options that don't involve using the VMM/App Controller interface (i.e., calling the PowerShell cmdlets directly).
    William Busby, PMP

    Nope.  not at this time.
    VMM focuses on managing the fabric components that the VMs require.
    You can build a Service Template that stands up an app server as a VM.  But, it would be managed by VMM.  So it really does not fit your ask.
    You need to look into configurations and desired state configuration.  At least that way you can define your end result and apply it to any machine.  But you still have to automate the deployment of the OS somehow.  Using WDS if you want to
    go that route, or any of the 3rd party management layers.
    Brian Ehlert
    http://ITProctology.blogspot.com
    Learn. Apply. Repeat.

  • Disk Utility in Lion 10.7.4 for Bare Metal Disaster Recovery??

    We just got our 15" Macbook Pro back from repair of a catastrophic hard disc failure. Wow. An older Seagate Momentus XT 500 toasted its little self. It seems most of our data are intact. If any software on the *new* hard disc get corrupted, we need a fairly straightforward disaster recovery process.
    I need to ask if anyone has successfully used Disk Utility to create a (bare metal) disaster recovery image of everything on their hard drive, including all their software *and* the perfectly functioning OS. We'd like to be able to restore a complete disc image of the new Seagate Momentus 750 XT - taken from a 1 TB Seagate FreeAgent GoFlex USB drive - to the Macbook.
    Has anyone done this successfully? Thanks very much for your help!

    Time Machine, SuperDuper and Carbon Copy Cloner are some of the utilities that will allow you to backup your disk and do a complete restore.

  • Can i run Mac OS X on Mac Mini under a "bare metal" Hypervisor?

    Unfortunately, I can't leave Windows because of heavy amount of legacy software which cannot be easily transfered to Mac OS X (for example, databases running on MSSQL).
    On the other side I like Mac Mini hardware for its size, noise level and power consumption and want to try Mac OS X (buying it only to replace Max OS X with Windows would be insane).
    So, to avoid buying a separate computer, I want to run both operating systems simultaneously under a "bare metal" Hypervisor (VMWare ESXi, Hyper-V etc.).
    Using BootCamp to run either Mac OS X or Windows doesn't fit my needs. Running Windows as a guest OS in Parallels - doesn't fit them too.
    Both OS should be able to run at the same time, and Windows should be on 24h a day.
    I know that a hypervisor can be installed on Mac Mini. I know also that running Windows on it won't be a problem.
    The only question is - will the Mac OS X Lion run under hypervisor?

    VMWare ESXi should support Lion and Mountain Lion, both in terms of it having been done and beeing allowed by Apple's license.
    Note: Apple's license only allows hosting two virtual OS X machines on a single host.
    In addition to ESXi, there is also Parallels Server for Mac Bare Metal Edition. See http://www.parallels.com/uk/products/server/mac/baremetal/technical-resources/ I merely mention it for completeness, I would stick to ESXi since it is 'free'.
    You could consider running OS X as the host operating system and Windows as a guest VM using VirtualBox. VirtualBox can be configured to start automatically and invisibly on the Mac host. You could then use Microsoft RDP to connect to the running VM. Its services like SQL Server would of course be running and accessible as normal.

  • Bare metal restore of full server backup fails with RPC error 0x800706BA (server 2012)

    With Windows Server 2012, after creating a successful windows server backup (full server), I am unable to do a bare metal restore of the system.
    EFI System Partition, C:, D: are the volumes on a single array (Raid 5).
    Loading WinRE and launching re-image, the full server restore stops after completing the C: restore just after beginning the D: drive restore (data volume) with the error with the message about RPC error 0x800706BA, unavailable.
    The system will not boot as a result of the failure. 
    What may be causing this?
    [edit]
    I think I have two things working against me here... one, Volume Shadow Copy service wasn't running in WinPE and two.. the D: volume contains data deduplication.
    I think this is causing some of my problems.
    Is there an additional service that I need to start in WinPE to get a volume that was deduplicated restored perhaps using wbadmin?

    Hi,
    Thanks for the reply Shaon.
    The hardware system is identical; however, I am increasing storage capacity from 12TB to 16TB.  Basically, the system is the same, just new hard drives (4x4TB versus 4x3TB).  Drives are all identical enterprise class drives and I am actually using
    the same system (backup, swap drives then restore).
    I tried both using the WinRE restore (re-image) with the backup set, and tried using wbadmin.  Both produce same RPC unavailable messages (although using the GUI based tool via WinRE provides the additional error code #).
    I do not have any idea what to look at to try to figure this one out.
    [edit]
    I was just informed of this posting:
    http://social.technet.microsoft.com/Forums/windowsserver/en-US/147efccc-e0e9-4458-852e-c3a44ae5cb64/the-system-image-restore-failed-error-details-he-rpc-server-is-unavailable-0x800706ba?forum=windowsbackup
    and in my case, this lab server does in fact act also as a DC as well as a file server...  do I need to spin up a secondary controller (perhaps in a VM on another machine) to do this restore?

  • Is WS2012R2E Capable of Bare Metal Restore to UEFI Client?

    Hi,
    After two weeks of fruitless effort attempting a client bare metal restore to a UEFI machine, I'm now wondering if it is supported in W2012R2E?
    This same machine is able to use WDS over the network to do a remote software install. Similarly, two other non-EUFI (BIOS) machines on the same network/server are able to do bare metal restores with no problem.
    No matter how the client UEFI machine is configured, it always gets to the same point in the "Full System Restore Wizard" and gets hung up - it can't find the server on the network. (Again, it works fine when using the same WDS server to do a remote
    software install - finds the server just fine).
    Any help/ideas would be greatly appreciated.
    Jeff

    Hi Justin,
    Let me take your questions one by one:
    Was it an error message when the Full System Restore Wizard hung up? Or no any error message, just the wizard hung up?
    What happens is that the Full System Restore Wizard reports that it cannot find the server. It then offers an option to either search again or
    to manually enter an IP address. When I enter an IP address manually, that's when the "Unknown Network Error" occurs. Note that repeated searches don't help. The Wizard never finds the server. See below where I give the full sequence of events.
    When run Full System Restore Wizard, please ping the server with server name and IP address from the client computer. Any find?I
    don't know that this is possible. At the point when the wizard is running, it's running under a downloaded image from the server (i.e. there IS connectivity to the server), but there's no OS to go to for the ping command. (If you know of a way to do that,
    please let me know and I will try it).
    The sequence of events is:
    The client computer is set up to boot via PXE over the network.
    The client computer receives a small net boot file. A text screen which offers the option to select either an x86 or X64 Full System Restore (I've tried both - doesn't make any difference). 
    Selecting one of the option results in another file downloaded from the server. Control is transferred to this file, and the display goes into graphical mode.
    A dialog box pops up allowing one to select language. Once that is selected, another window opens that shows progress. At some point in this process, the program indicates that it cannot find network drivers, and they need to be loaded via USB. (Note that,
    as a test, I've actually used WDS to load drivers into the image that was downloaded at this point - but it still claims it cannot find network drivers).
    A dialog box opens giving the opportunity to click to load drivers.
    I click the button. Another dialog opens and tells me to insert a USB key with drivers on it.
    I insert a USB key with network drivers. Click the button that says scan for network drivers.
    After a few seconds, the drivers are found and a dialog reports that fact.
    At that point, the window that appears is the Full System Restore Wizard. It asks to click "Next" to search for Windows Servers.
    When I click next, the cursor changes to a "waiting" cursor for about 30 seconds, while presumably the wizard is searching for servers. (Note that the machine just completed downloading at least two files from that very server).
    After some time elapses, a dialog appears indicating that the server couldn't be found. It offers two options, "Search Again", or manually enter a server name or IP.
    Searching again has the same results. Manually entering an IP or server name results in the "Unknown Network Error" I've referred to.
    At this point, you cannot proceed further. Clicking "Next" causes the network error message to come back again. Clicking "Cancel" drops you out of the wizard and the computer reboots.
    Please check those log files which were modified when run the Full System Restore Wizard.
    I will try this again when I arrive home tonight and examine the logs.
    Thanks,
    Jeff

  • What is your preferred method of remote bare metal hyper-v 2012 server imaging

    I need to remotely provision (image) 32 Dell 620 servers (they have iDRAC installed), I will have system center 2012. I was thinking PXE boot and place images on the servers. I know I can use the iDRAC to remote boot a virtual ISO file and install the OS
    that way. I am looking for a better faster and more pre-configured OS way.
    I am used to using Shadow Protect (ghost like product) and making and placing 0 and 1 disk images on servers, I would like to continue along that vein. But with added ability to PXE boot and image the server that way. I am looking into clonezilla and want
    any thoughts of success or failures with server 2012 deployment.
    I understand that you can use system center 2012 to deploy images to your bare metal servers (lots of moving parts and looks pretty complicated) and that it will then deploy a VHD image file to your server. I am assuming that means it will boot to the VHD
    file (hyper-v light) . I have heard this slows down disk read\writes and can cause some issues.
    Thoughts with clear instructions please. Only link me to a website if it helps you to not type so much.
    Also looking into the DELL KACE server.
    Can anyone tell me if I need system center only or if I need Configuration Manager to image servers??
    [email protected] visit my blog stephanco.blogspot.com

    You can use SCVMM 2012/2012R2 bare metal deployment
    http://blogs.technet.com/b/cloudytom/archive/2012/01/02/scvmm-2012-bare-metal-deployment-checklist.aspx
    Darshana Jayathilake

  • Meeting Place 8.5 install on Bare Metal UCS-C

    I am trying to install MP 8.5, Webex Scheduling with Two Application Server nodes.
    One Application Server is to be put on VM 5.0 on a UCS_C.
    The second Application Server is to be put directly on a Bare Metal UCS_C, NO VMware.
    Is the second scenario supported by Cisco and TAC?   
    thanks
    Any help will be appreciated.  

    Hello tknapp,
    You will need WMware for MP, keep in mind that MP can be the only VM running on a server (at east on C210M2). My understanding is that Cisco only supports UC apps running on  top of a WMWare hypervisor on UCS servers.
    You will have to pay for 2 CPU license to VMware for each server. So in essence MP (in it's present form) can be virtualized but not really, as it can be the only VM on the physical box.
    Regards,
    Greg.

  • Bare-metal deployment to previously used host fails - removing host references.

    I am running SCVMM 2012 R2 and setting up bare-metal deployment of Hyper-V hosts. I got it almost figured out but I am getting a failure when it tries to register the host with the error The fault reason was: 'Unexpected registration of physical
    computer 4a310166-642f-e211-0050-05000000003a.' in the vmm log file vmmagentpe.exe.log. The error onscreen in PE is
    Error: 803d0013. I have been using this server for my testing and have had a few other failures which are now resolved. All other tasks in the job complete except for
    Wait for physical machine to reboot and customization to be finished.
    This host was also registered in VMM as part of a manually built cluster before I started looking into bare-metal. I believe the problem is that there are probably references to this host in VMM.
    Does anyone know where to look in the db or another method to remove anything related to this host?

    I figured this out. It was rebooting to PXE instead of booting to the newly installed OS. once I reset the boot order to local disk it came up and ran mini-setup, domain join and all as expected.
    -tim

  • Bare Metal Deployment

    When Deploying a VHD through Bare Metal you end up with a C: that is 40 BG is size and a D: the size of the physical drive. Is there any way to change the size of the resulting C:
    Is there any good documentation on this

    The size of the "C:" drive is dependant on the size of the VHD(x) file you are deploying to the host.
    You will always see the D drive and it will always show the entire size of the physical drive as it is indeed the physical drive that hosts the VHD(x) file (the C drive is the mounted VHDx file stored on the "D" physical drive).
    If you want to see the C drive as using all of the available space the physical drive has, size it appropriately when creating it.
    However... 50-60Gb should be sufficient for any deployment as it should only ever be the Base OS, plus Hyper-V component + drivers + Management tools/Agents. The physical drive needs enough space to store the expanded size of the VHD(x) file plus the Page
    File (this is created outside of the VHD(x) file).
    This also means that you can then use that VHD(x) file to deploy to any server with a physical drive bigger than say 60Gb rather than sizing the VHD(x) file at 200Gb and only being able to use it with servers with 200Gb+ physical drives.
    This link might help you understand it more as it's a general Windows feature not VMM/Hyper-V:
    http://technet.microsoft.com/en-us/library/hh825689.aspx
    My System Center Blog

Maybe you are looking for