Best Practices in OS imaging and Deployment

PC imaging

  1. Check for power source:

    Ensure that the system where imaging is performed is connected to a power source. This is needed since imaging might take a longer time and if the system battery is low during imaging, it might lead to disk error. 

  2. Disk Defragmentation:

    Defragment the disks and image repository drive before initiating image creation. Defragmenting will pull all the sectors of the hard disk together. This will increase the disk access which will in turn lead to faster Image Creation. Follow the steps given below to defragment disk,

    • Go to Control panel --> System and Security --> Administrative tools
    • Select Defragment and Optimize Drives
    • Select the drives that you want to defragment and click Analyze
    • Once the drives are analyzed, click Defragment
  3. Check for Disk bad sectors:

    • Open command prompt with administrative privileges and run command "chkdsk <drive letter:> /f" (for eg, "chkdsk d: /f") to check the bad sectors in your OS drive. This fixes the logical file system errors that occur in volume bitmap and ensures that the drive is error free. 
    • You can also locate the bad sectors and recover readable information by using command "chkdsk <drive letter:> /r" (for eg, "chkdsk d: /r"). However, this will take hours to recover the information. 
  4. Check for patch updates and software installation:

    Do not initiate image creation process if any patch updates or software installation is running in the system, since partial updates might lead to boot errors after deployment.

  5. Configuring repositories:

    While creating Image and driver repositories, you can specify the computer name in share path if DNS is configured in your network. If it is not configured, you can specify the IP addresss to avoid network errors. For networks with dynamic IP, make sure that you change the IP address when the PC is moved to another location. This is needed since deployment cannot happen if driver or image repository is not accessible. 

  6. Check for Disk Usage:

    Do not initiate online image creation if the disk usage for any process in the machine is high. You can find the details of disk usage in task manager. For example, if the disk usage is above 400 MB/s you can either wait for the process to complete or kill the process before starting image creation.

  7. Naming and renaming Image file or repository:

    Give unique and easily identifiable names for images, templates, repositories, and media. After creating images, do not rename image repository or the image file. If the image file or the repository is not accessible because of the changes, deployment cannot be initiated.

  8. Memory level (RAM):

    Ensure that you select the required memory level (RAM) during imaging. It is recommended to select the memory level as medium, since, if the memory level is low, image creation speed might slow down and if the memory level is high, too much of RAM will be consumed during the imaging process. If the imaging machine is going to be idle, you can select memory level as High, since the RAM consumption will not affect any other process and the imaging will be faster. 

  9. Configuring driver repository:

    During the process of imaging, drivers will be automatically collected from the imaging machine and stored in a repository. It is recommended to use a local share while configuring this repository, since this repository is needed to be accessible always to store the drivers from the imaging machine.

  10. Check for OS partitions:

    During imaging, System/Firmware partitions, OS reserved partitions and OS partitions will be selected as default, ensure that you do not deselect them, as these partitions are required to boot the system after OS deployment.

  11. Check for external devices:

    Ensure that you do not connect or disconnect any external devices to the machine during the imaging process. When it is connected, the newly detected registry components might interrupt the imaging process.

PE media creation 

  1. Drivers Re-scan:

    Network and hard disk drivers are needed for PE media creation. Hence, once you add these drivers to the repository get them updated using the Re-scan option for the changes to reflect. 

  2. PXE server port:

    Ensure that port 69 is not used by any other process as it is used for configuring PXE server.  Refer this document for more details 

    To know more about the ports used in OS Deployer, refer this document.

  3. Hard disk and Network drivers:

    Ensure that you have added the compatible network and hard disk drivers in the driver repositories to create PE media. These drivers should support the WinPE toolkit and the target computers used for deployment.  

OS Deployment

  1. Check for power source:

    Ensure that the target computer where deployment is performed is connected to a power source. This is needed since deployment might take a longer time and if the system battery is low during deployment it might lead to disk error. 

  2. Provide Administrator privileges to user for post deployment activities:

    During template creation, while adding the domain details, give administrators privileges to the credentials to perform post-deployment activities.

  3. Hard disk number:

    While creating deployent templates, select the appropriate hard disk number to perform the deployment and make sure that the target computer have the required hard disk drivers to perform deployment. In general, the value of default internal disk will be 0. Open command prompt, and run command "diskpart" and "listdisk" to know hard disk number. 

  4. Mac Address:

    If you choose to deploy by specifying Mac address, manually boot the computers to initiate deployment. To automate the booting process, configure wake on LAN settings in all the target computers. 

  5. Bandwidth level:

    You can restrict the usage of bandwidth in your network by setting up the bandwidth usage level. This setting is needed to be configured, especially when you use multicast method since more bandwidth will be consumed while deploying the same image to multiple computers at a time. 

  6. Waiting time for deployment:

    Specify the required waiting time before initiating deployment. During this time the server will wait for the target computers to connect.

    • In case of multicast deployment, the server will wait for the computers to connect in this time interval and deployment will be initiated only after the configured waiting time. This is because the same image copy will be sent to all the computers at a time. Ensure that the wait time you configure is sufficient for the required computers to connect.
    • In case of unicast deployment, the deployment will be initiated as soon as the computer connects to the server, since image copy will be deployed to each computer separately. Deployment will be initiated only when the computer connects to the server within the specified time interval.