Target Servers With Multiple Disks Deployment Support
Article: 20000038
Applies to: Release 6.x
Related Articles
Overview
Configuring Scripted Installs
Configuring Image Installs
Related Articles
Overview
By default, the Create Disk Image and Distribute Disk Image tasks operate on the "first" disk.
The disks are enumerated using a proprietary algorithm. The "first" disk might not be the system boot disk.
The Altiris showdisk utility can be used to generate human-readable disk enumeration output. This utility is
called in the provided Get Drive Configuration jobs.
Configuring Scripted Installs
The scripted install jobs use a small image to partition and format the disk. When deploying to a server with multiple disks,
either DAS or SAN, the system boot disk might need to be specified in the job for the NTFS.img or GRUB.img images.
To specify a particular disk for all other operating systems:
| 1. |
In the Deployment Console, copy and rename the scripted install job. In the new job, edit the
Distribute Disk Image task with adding the -dx switch to the Additional Parameters field
where x is an integer that represents the appropriate disk number.
|
| 2. |
For Windows Server 2008 and 2008 R2, the unattend file must specify the system boot disk.
| a. |
On the Deployment Server, copy and rename the unattend file. Modify the <DiskId> field.
|
| b. |
In the new job from step 1, edit the Run Script - Configure Answer File task to use the new unattend file.
|
|
| 3. |
For VMware ESX 4.x or Red Hat Linux, the kickstart file must specify the system boot disk.
| a. |
On the Deployment Server, copy and rename the kickstart file.
Modify the new kickstart file by replacing the autopart command with separate part
command lines and adding the --ondisk=XXX option to each where XXX is the device label.
Example for non-LVM:
part /boot --size=75 --ondisk=cciss/c0d0
part swap -size=1000 --ondisk=cciss/c0d0
part / --size 5120 --grow --ondisk=cciss/c0d0
Example for LVM:
part /boot --size=100 --fstype=ext3 --ondisk=cciss/c0d0
part pv.01 --size=100 --grow --ondisk=cciss/c0d0
volgroup VolGroup00 --pesize=32768 pv.01
logvol swap --fstype=swap --name=LogVol01 --vgname=VolGroup00 --size=1000 --maxsize=1984
|
| b. |
In the new job from step 1, edit the Run Script - Configure Answer File task to use the new kickstart file
for the export basekickstartfile= value.
|
|
Configuring Image Installs
When capturing or deploying an image to a server with multiple disks, either DAS or SAN, the system
boot disk in the job might need to be specified.
To specify a particular disk:
| 1. |
In the Deployment Console, copy and rename the read and write job pair. Edit the
Create Disk Image or Distribute Disk Image task with adding the
-dx switch in the Additional Parameters field
where x is an integer that represents the appropriate disk number.
When deploying a Linux image with LVM, x must be a comma-separated
list of integers representing all of the disk numbers that the LVM volume will span.
For example, to deploy an LVM image that spans the first three disks, use -d1,2,3.
|
| 2. |
To capture or deploy additional drives on target servers running Windows, add a Create Disk Image
or Distribute Disk Image task immediately after the modified image task from step 1.
| a. |
Use a unique disk image filename. |
| b. |
Add the rdeploy command line switch, -dx, where x is the drive number. |
| c. |
For captures, select Do not boot to Windows or Do not boot to Production to prevent
unnecessary reboots.
For deploys, modify all Distribute Disk Image tasks, except the
last drive, to have the Automatically perform configuration task after completing this imaging task
not selected. Only the Distribute Disk Image task for the last drive should have this checkbox selected.
|
| d. |
Select the appropriate Automation pre-boot environment. |
| e. |
Do not select Prepare using Sysprep or Prepared using Sysprep for data drives. |
|
IMPORTANT: If one of the logical drives on the target server is designated the "F"
drive, the image capture will fail with the error, "System cannot find the file specified."
This error is explained in the Altiris Knowledgebase (Article 33065).
An alternate solution is to rename an existing "F" drive with another designation.
IMPORTANT:Deploying a Red Hat Linux non-LVM image could lead to an "Invalid Partition" error
when the target server attempts to boot. This is caused by the reference server from where the image was
captured had the Linux partitions installed across all drives; however, the image file captured
only contained partitions from one of the disks. The image file must be captured either from a target
with Red Hat Linux and LVM or where all Linux partitions are forced onto the first drive. The later
can be accomplished by modifying the kickstart file as described in the Configuring Scripted Installs section.
|
Printable version
