Deployment Guide

Q: When I use the sample DTK scripts to deploy multiple systems, do I edit the configuration files to reflect unique information
(such as unique system names, IP address, and BIOS asset tags) for each system?
A: The sample scripts are provided as examples for users who want to develop their own deployment process. The scripts may work
perfectly in your environment. If not, you may need to develop your own scripts entirely from scratch. If you are deploying multiple
systems, for example, provide unique information for each system when appropriate. To perform this task, modify each .ini (or other
configuration script) file to reflect the unique information for each system you are deploying (such as the remote access controller
(RAC) IP addresses and BIOS asset tags). There are many options available to optimize this process and it is suggested that you
conduct an Internet search for available tools.
Q: When I change my hard drive controller from SCSI to RAID (or RAID to SCSI), the system prompts me for confirmation
during POST. How can I stop this from happening?
A: Use the --noraidprompt option with the --embscsiraid option to prevent the system from prompting during POST. Data loss
results from changing the state of the disk controller, so you must be certain before skipping the prompt. There is no method for
using the ‑‑noraidprompt in the input file. If you want to change the controller state through an input file, you must configure your
script to call the SYSCFG.EXE utility twice, once to specify the input file, and once to change the controller state with the --
noraidprompt option.
Q: What do I do if the PARTCFG script fails?
A: Ensure that DT_HD is set to the required disk to be configured before you execute the PARTCFG script.
RAIDCFG questions
Q: Why does RAIDCFG display an error message when I use an invalid read, write, and cache policies and/or stripe sizes for a
particular controller?
A: All RAID controllers have their own default read, write, and cache policies and stripe sizes. It is possible that if you provide an
invalid policy or a stripe size for a particular controller, RAIDCFG may not give an error but creates the virtual disk with the default
policy and/or stripe size.
Q: When I execute RAIDCFG with the RAID level of "01," RAIDCFG creates RAID 1, not RAID 01.
A: This behavior occurs because the RAIDCFG CLI parser ignores zeros preceding any parameter value. For example, -r=01 is
parsed as -r=1.
Q: What is strict creation?
A: Strict creation is an optional flag that has been added to the create virtual disk command. It allows you to create virtual disks
only if the array disks (that is, hard drives) are within a specified percentage of disk space of each other.
Q: Why is the drive location of array disks displayed like 0:0:0?
A: The nexus used to display array disks is "channel:target:lun." The Logical Unit Number (LUN) is always "0." The channel and the
target IDs are necessary to identify array disks when dealing with SCSI devices.
When dealing with SAS devices, the array disk location is displayed as channel:target:enclosure. It is possible to have the enclosure
value as non-zero. If the enclosure has a non-zero value, then all three numbers (channel, target, and enclosure) have to be included
on the command line.
Q: Do I need to specify RAID type and size when creating virtual disks?
A: No. The default RAID type is RAID 0. If RAID size is not provided, the maximum size allowed for virtual disk is created. The
mandatory fields required to create virtual disks are the controller slot ID and the array disks.
Q: What is a hot spare (failover drive)?
65