Creating SDS Guests from OVA (KVM or OpenStack)

2.3 Creating SDS Guests from OVA (KVM or OpenStack)

Perform the following procedure in Cloud Client to create SDS Guests from OVA (KVM or OpenStack):

Pre-requisites:

Create instance flavors.
Use the DSR Cloud Benchmarking Guide values to create flavors for each type of VM. Flavors can be created with the Horizon GUI in the Admin section or the nova flavor-create command line tool.
Name the flavors describing their respective function and ensure it is as informative as possible. Flavors define resource sizing. A standard convention is to name as 0406060. Where the first two digits (04) represent the number of virtual CPUs, the next two digits (06) represent the RAM allocation in GB, and the final three digits (060) represent the disk space in GB.
The default Large Receive Offload (LRO) option must be disabled on the host command line while using an Intel 10 Gigabit Ethernet ixgbe driver on the host nodes. For more information, see the Intel release notes.
```
$ sudo ethtool -K <ETH_DEV> lro off
```

To add SDS OVA image, perform the following steps:
1. Copy the OVA file to the OpenStack control node.
```
$ scp SDS-x.x.x.ova admusr@node:~
```
2. Log into the OpenStack control node.
```
$ ssh admusr@node
```
3. Unpack the OVA file using tar in an empty directory.
```
$ tar xvf SDS-x.x.x.ova
```
4. Import the unpack the .vmdk VM image file.
```
SDS-x.x.x-disk1.vmdk
```
5. Source the OpenStack admin user credentials.
```
$ . keystonerc_admin
```
6. Select an informative name for the new image.
```
sds-x.x.x-original
```
7. To use VMDK format, import the image using the glance utility from the command line.
```
$ glance image-create --name sds-x.x.x-original --is-public True --is-protected False --progress --container-format bare --disk-format vmdk --file SDS-x.x.x-disk1.vmdk
```
  Note:
  This process takes about five minutes depending on the underlying infrastructure. To complete the VMDK format, go to Step 2.
8. To use QCOW2 format, perform the following steps:
  1. Convert VMDK to QCOW2 format using the qemu-img tool and create a qcow2 image file running the following command:
```
qemu-img convert -f vmdk -O qcow2 <VMDK filename> <QCOW2 filename>
```
    For example:
```
qemu-img convert -f vmdk -O qcow2 SDS-82_12_0.vmdk SDS-82_12_0.qcow2
```
    Note:
    If the qemu-img tool is not already installed, install it using yum command.
    sudo yum install qemu-img
  2. Import the converted QCOW2 image using the glance utility from the command line.
```
$ glance image-create --name sds-x.x.x-original --is-public True --is-protected False --progress --container-format bare --disk-format qcow2 --file SDS-x.x.x-disk1.qcow2
```
    Note:
    This process takes about five minutes depending on the underlying infrastructure.
To name the new VM instance, perform the following steps:
1. Create an informative name for the new instance.
```
SDS-NOAM-A
```
2. Review the network interface recommendation provided in DSR Cloud Benchmarking Guide.
In OpenStack Control Node, create and boot the VM instance from the glance image by performing the following steps:
1. Retrieve the required configuration values, by running the following commands:
  - To obtain the image ID:
```
$ glance image-list
```
    Output:
```
811f0181-6e66-4cf0-9eb7-8058d86edf05
```
  - To obtain flavor ID:
```
$ nova flavor-list
```
  - To obtain network ID(s):
```
$ neutron net-list
```
    Output:
```
cb2a0b22-2383-462d-bce5-73f3f5bb752d
```
2. Specify what information should convey by its name.
```
SDS-NOAM-A
SDS-NOAM-B
```
3. Create and boot the VM instance.
  
  Note:
  The instance must be owned by the DSR tenant user, not the admin user.
4. Source the credentials of the DSR tenant user and run the following command:
```
$ nova boot --image <image ID> --flavor <flavor id> --nic net-id=<first network id>,v4-fixed-ip=<first ip address> --nic net-id=<second network id>,v4-fixed-ip=<second ip address> -–config-drive true <instance name>
```
  Note:
  Use one --nic argument for each IP or interface.
  IPv6 addresses should use the v6-fixed-ip argument, instead of v4-fixed-ip.
5. To verify if the new instance has been booted, view the newly created instance using the nova tool.
```
$ nova list | grep -i (xmi address)
```
  Note:
  The VM takes approximately five minutes to boot and can be accessed through both network interfaces and the Horizon console tool.
To configure VIP, in OpenStack Control Node, perform this step.
Note:
This is an optional step. Refer to Application VIP Failover Options (OpenStack) for more information on VIP.
If a NOAM or SOAM VIP is needed, run the following commands:
1. Find the port ID associated with the instance’s network interface.
```
$ neutron port-list
```
  Output:
```
aed2522e-cf52-4aa4-9e12-4acab7f8df04
```
2. Add the VIP IP address to the address pairs list of the instance’s network interface port.
```
$ neutron port-update <Port ID> --allowed_address_pairs list=true type=dict ip_address=<VIP address to be added>
```
Check if interface is configured.
If DHCP is enabled on Neutron subnet, VM configures the VNIC with the IP address provided in Step 3. To verify, ping the XMI IP address provided with the nova boot command from Step 3:
```
$ ping <XMI-IP-Provided-During-Nova-Boot>
```
If the ping is successful, ignore Step 6 to configure the interface manually.
To manually configure interface, in OpenStack Dashboard (Horizon), perform the following steps:
Note:
This is an optional step.
If the instance is already configured with an interface and has successfully pinged in Step 5, then ignore this step to configure the interface manually.
1. Log into the Horizon GUI as the tenant user.
2. Navigate to the Compute/Instances section.
3. Click Name field of the newly created instance.
4. Select Console tab.
5. Login as the admusr user.
6. Select an informative hostname for the new VM instance.
```
SDS-NOAM-A
SDS-SO2
```
7. Configure the network interfaces, by conforming to the OCDSR Network to Device Assignments defined in the DSR Cloud Benchmarking Guide.
```
$ sudo netAdm set --onboot=yes --device=eth0 --address=<xmi port ip> --netmask=<xmi net mask>
$ sudo netAdm add --route=default --device=eth0 --gateway=<xmi gateway ip>
```
  Note:
  Under some circumstances, it may be necessary to configure more interfaces.
8. If netAdm fails to create the new interface (ethX) because it already exists in a partially configured state, run the following commands:
```
$ cd /etc/sysconfig/network-scripts
$ sudo mv ifcfg-ethX /tmp
```
  Note:
  Keep ifcfg-ethX in /tmp until ethX is working correctly.
9. To create and configure the interface in one action, re-run the netAdm command.
10. Reboot the VM.
```
$ sudo init 6
```
  Note:
  It takes approximately five minutes for the VM to complete rebooting.
  
  The new VM should now be accessible using both network and Horizon console.
Verify network connectivity, by pinging the default gateway.
```
$ ping –c3 <gateway address for the External management network>
```
Note:
Repeat these Step 2 through Step 7 for each server before continuing to the next procedure (for example, NOAM-A, NOAM-B, DR Servers, Query Server, and DP).