Strategies for Replicating HP-UX 11i Virtual Systems onto Physical and Virtual Integrity Servers

You need only execute this command once after installing the device on the physical VM Host system. This
creates device special files in /dev/rscsi enabling Integrity VM’s attached I/O functionality. Suppose that
the physical VM Host system has a tape drive with device special filename /dev/rmt/c6t1d0 that your
virtual machine requires access to. The SCSI pass-through file associated with /dev/rmt/c6t1d0 is
/dev/rscsi/c6t1d0 and you may attach the tape drive to your VM by executing:
# hpvmmodify –P mynewvm –a tape:scsi::attach:/dev/rscsi/c6t1d0
Virtual Device Identification and Removal
When a virtual machine no longer needs connectivity to a storage device, detach the device by using the
Integrity VM command line interface and explicitly identifying the virtual device. The explicit definitions are
most easily identified by using the hpvmstatus command with the d (device listing) option. For example:
# hpvmstatus –P mynewvm -d
[Virtual Machine Devices]
[Storage Interface Details]
disk:scsi:0,0,0:file:/hpvm/VHD/vmhpux0.disk0
disk:scsi:0,0,1:disk:/dev/rdsk/c4t12d0
disk:scsi:0,0,2:disk:/dev/rdsk/c5t2d5
tape:scsi:0,0,3:attach:/dev/rscsi/c6t1d0
If you need to detach the tape drive from your VM, then use the –d option of hpvmmodify and the explicit
device specification as follows:
# hpvmmodify –P mynewvm –d tape:scsi:0,0,3:attach:/dev/rscsi/c6t1d0
When removing virtual disk drives, take care to unmount the file system associated with the disk, reconfigure
any logical volume groups associated with it, etc. before detaching it from the VM.
Planning the Deployment
Before cloning the virtual system and replicating it on other systems, take care to consider the following:
Identify the virtual hardware configuration of the source virtual system (with hpvmstatus –P <vmname>) and
make sure the target system’s hardware is configured with computer resources sufficient for deploying the
source system. That is, the target system should have memory and mass storage capacity sufficient to contain
the deployed system. The target system should have necessary CPU and network resources as well. Target
systems with a “weaker” hardware configuration may encounter performance problems.
If you made specific system tuning changes (beyond defaults) to the source system, then those changes need
to be revisited. For example, the maximum number of processes (nprocs) may change depending on CPU
core count. Similarly, file system buffer cache size is defined as a percentage of physical memory and may
need reconsideration. Keep in mind that system installation configures tunables based on hardware
configuration.
Identify the device drivers needed for the target system’s storage and network hardware, installing them on
the source system if necessary – before cloning the source system. Fibre channel device drivers are a good
example of device drivers that may need to be manually installed on the source system as they may be
required for the system’s boot disk.
Verify that the software installed on the source system is licensed for use on the target systems on which it is
being replicated.