Check host station OS compatibility
|
|
|
1 |
Check the NSP Release Notice to ensure that the OS version of the host station supports the creation of VMs at the RHEL version that the NSP requires.
|
2 |
Log in to the VM host station as the root user.
|
3 |
If the host station OS version supports NSP VM creation, enter the following; otherwise, update the host OS version as required:
# osinfo-query os | grep rhel | grep -v - ↵
A list of supported RHEL variants is listed, for example:
rhel7.8 | Red Hat Enterprise Linux 7.8 | 7.8 | http://redhat.com/rhel/7.8
rhel7.9 | Red Hat Enterprise Linux 7.9 | 7.9 | http://redhat.com/rhel/7.9
rhel8.0 | Red Hat Enterprise Linux 8.0 | 8.0 | http://redhat.com/rhel/8.0
rhel8.1 | Red Hat Enterprise Linux 8.1 | 8.1 | http://redhat.com/rhel/8.1
rhel8.2 | Red Hat Enterprise Linux 8.2 | 8.2 | http://redhat.com/rhel/8.2
|
4 |
Record the appropriate RHEL version number in the left column, which is one of the following:
-
the version that matches the NSP-supported RHEL version, if listed
-
the version that is less than but closest to the supported NSP RHEL version; in the output example, the version to record is 8.2, as the NSP supports a higher RHEL version that is not listed
|
Prepare required images
|
|
|
5 |
Log in to the host station as the root user.
|
6 |
Download one of the following files from the NSP downloads page in the OEM_Images subdirectory on the Nokia Support portal to a local directory on the station:
-
NSP_K8S_PLATFORM_RHEL8_yy_mm.qcow2—for NSP deployer host or NSP cluster VM
-
NSP_RHEL8_yy_mm.qcow2—for components outside the NSP cluster, such as an NSP auxiliary databases, and NFM-P components
where yy_mm represents the year and month of issue
|
7 |
Open a console window.
|
8 |
Enter the following:
# dnf -y install virt-install libguestfs-tools ↵
|
9 |
For each VM that you require, enter the following to create a raw VM disk image file:
# qemu-img convert -f qcow2 qcow2_file -O raw -S 0 raw_image.img ↵
where
qcow2_file is the name of the downloaded qcow2 file
raw_image is the name that you want to assign to the image; for example, NSP_Server_A
|
10 |
Perform one of the following:
-
If you want only one disk to contain all OS, product software, and data files on a VM, you must resize the VM disk image in accordance with the response to your Platform Sizing Request.
For each one-disk VM that you require, enter the following:
# qemu-img resize -f raw "raw_image.img" sizeG ↵
where
raw_image is the raw disk image name specified in
Step 9
size is the required disk size, in GBytes
-
If you want more than one disk in a VM, for example, one for the OS, and one for all NSP component software and data, or separate disks for specific partitions, you must create a separate raw image for each required disk. The disk size must be in accordance with the response to your Platform Sizing Request.
For each separate disk image that you require, enter the following:
# qemu-img create -f raw "raw_image.img" sizeG ↵
where
raw_image is the name that you want to assign to the disk image; for example, NSP_Server_A_Complete, for an image that is to contain all NSP component server partitions, or NSP_Server_A_Software, for an image that is to contain only the /opt/nsp partition
size is the required disk size, in GBytes
|
11 |
The raw image files that you create in
Step 10 are in sparse format; you must convert the image to non-sparse format, which provides optimal disk performance.
Perform the following steps for each raw disk image created in
Step 10.
-
Enter the following:
# cp --sparse=never raw_image.img non-sparse_image.img ↵
raw_image is the name of a raw disk image created in
Step 10
non-sparse_image is the name to assign to the non-sparse image
A non-sparse_image.img file is created.
-
Delete the raw_image.img file, which is no longer required.
|
Deploy VMs
|
|
|
12 |
Enter the following once for each VM to deploy the VM:
Note: One “--network bridge=bridge_name” entry is required for each VM interface that you intend to configure.
# virt-install --connect qemu:///system --ram RAM --vcpu=vCPUs -n instance --os-type=linux --os-variant=variant --disk path="image_name", device=disk,bus=virtio,format=raw,io=native,cache=none --network bridge=bridge_name --import & ↵
where
RAM is the required amount of VM RAM in the response to your Platform Sizing Request, in MBytes; for example, 64 GBytes is expressed as 65536, which is 64 x 1024 MBytes
vCPUs is the required number of vCPU threads in the response to your Platform Sizing Request
instance is the name to assign to the VM
variant is the OS version recorded in
Step 4, for example, 8.2
image_name is the name of the raw or non-sparse disk image created for the VM
bridge_name is the name of the network bridge for a VM interface
|
13 |
Enter the following to open a console session on the VM:
# virsh console VM ↵
where VM is the VM name
You are prompted for credentials.
|
14 |
Enter the following credentials:
A virtual serial console session opens on the VM.
|
15 |
Configure the RHEL OS as required for the NSP component; for example:
Note: If an NFM-P system integrated with the NSP uses hostnames, the NSP cluster VMs must be able to resolve each NFM-P hostname using DNS.
|
16 |
Perform one of the following; see
Disk partitioning for trial deployments or
Disk partitioning for live deployments for component-specific partitioning schemes.
Note: If you are using multiple disks in a VM, you must mount a parent partition before you mount any child partition. For example, you cannot mount the /var/log/audit partition before you mount the /var/log partition.
-
If you are using only one disk per VM, perform the following steps for each such VM.
-
Enter the following commands:
# mkdir -p /extra ↵
# mkdir -p /opt/nsp ↵
-
Use the RHEL fdisk utility to create the required sub-disks for the following directories:
For each directory, enter the following and then respond to the prompts; use the directory size value from the response to your Platform Sizing Request:
# fdisk /dev/virtual_device ↵
where virtual_device is the virtual device name, for example, vda in a KVM VM
-
Enter the following to reboot the VM:
# systemctl reboot ↵
-
After the reboot, perform one of the following.
-
If you are using LVM, perform the following steps.
1. Enter the following sequence of commands for each sub-disk:
# pvcreate /dev/virtual_devicen ↵
# vgcreate vg2 /dev/virtual_devicen ↵
where
virtual_device is the virtual device name, for example, vda in a KVM VM
n is the number associated with the sub-disk
2. Go to
Step 17.
-
If you are not using LVM, perform the following steps.
1. Enter the following for each sub-disk:
# mkfs fs_type -L path /dev/devicen ↵
where
fs_type is the file system type; see
Supported file systems for partition-specific file system support
path is the directory path associated with the sub-disk, for example, /opt/nsp
device is the device name, for example, vda in a KVM VM
n is the device number associated with the sub-disk
2. Open the /etc/fstab file using a plain-text editor such as vi.
3. Add one line in the following format for each sub-disk:
/dev/virtual_devicen path fs_type defaults 0 0
where
device is the device name, for example, vda in a KVM VM
n is the number associated with the sub-disk
path is the directory path associated with the sub-disk, for example, /opt/nsp
fs_type is the file system type; see
Supported file systems for partition-specific file system support
4. Save and close the file.
5. Enter the following:
# mount -a ↵
6. Go to
Step 19.
-
If you specify multiple disks per VM and are using LVM, enter the following sequence of commands for each disk in each VM:
# pvcreate /dev/device ↵
# vgcreate group /dev/device ↵
where
device is the device name for the disk
group is the name to assign to the volume group, and must be unique in the VM
|
Configure LVM
|
|
|
17 |
Create the LVM volumes and partitions.
Perform the following steps for each disk in a VM, beginning with the parent disk partitions.
Note: If you are using multiple disks in a VM, you must mount a parent partition before you mount any child partition. For example, you cannot mount the /opt/nsp/nfmp/nebackup partition before you mount the /opt/nsp partition.
Note: The /extra partition is allocated for use as a temporary storage location for downloaded product software.
-
Enter the following to create a logical volume:
# lvcreate -n volume -L sizeG group /dev/device ↵
where
volume is the name to assign to the logical volume
size is the required volume size in the response to your Platform Sizing Request
group is the name to assign to the volume group, and must be unique in the VM
device is the device name
-
Enter the following:
# mkdir directory ↵
where directory is the name of the directory to associate with the volume, for example, /opt/nsp
-
Enter the following:
# mkfs fs_type -L directory /dev/group/volume ↵
where
fs_type is the file system type; see
Supported file systems for partition-specific file system support
directory is the directory associated with the volume
group is the volume group
volume is the logical volume name
-
Open the /etc/fstab file using a plain-text editor such as vi.
-
Add an entry in the following format:
/dev/group/partition directory fs_type noatime 0 0
where
group is the volume group
partition is the partition name
directory is the associated directory path
fs_type is the file system type; see
Supported file systems for partition-specific file system support
-
Save and close the file.
-
Enter the following:
# mount -a ↵
|
Perform optional security hardening
|
|
|
18 |
Optionally, for greater system security, you can remove the world permissions from RHEL compiler executable files, as described in
Resetting GCC-compiler file permissions.
|
19 |
Close the open console windows.
End of steps |