ZStack Cloud arranges compute resources through zones,
clusters, and hosts. Therefore, the location of a VM instance can be identified
according to the zoneUuid, clusterUuid, and hostUuid.
When a VM instance is running, these UUIDs will be set to values that
represent the current location of the VM instance.
After a VM instance is stopped, hostUuid will be set to NULL, while zoneUuid
and clusterUuid will remain unchanged.
The lastHostUuid parameter is special, as it represents the host on which
the VM instance was running last time.
For newly created VM instances, lastHostUuid is set to NULL.
Once a VM instance is stopped, lastHostUuid is set to the previous value of
lastHostUuid.
Networks
A VM instance can have one or more L3 networks. The VM NIC Inventory section lists
the IP address, netmask, MAC address, and other information about each L3
network.
If a VM instance has more than one L3 network, you must set a default L3
network for this VM instance. The default L3 network provides default route,
DNS, and hostname.
If a VM instance has only one L3 network, this one automatically becomes the
default L3 network.
The following example helps you understand what a default L3 network is. Suppose you
have a user VM instance, as shown in Figure 1.Figure 1. User VM Instance
As shown in the figure above, the VM instance has three L3 networks, all of which
provide the Source Network Address Translation (SNAT) service. The default L3
network is
10.10.1.0/24.
CIDR: 10.10.1.0/24
Gateway: 10.10.1.1
DNS domain: web.tier.mycompany.com
Then, the routing table of the VM instance would be as
follows:
default via 10.10.1.1 dev eth0
10.10.1.0/24 dev eth0 proto kernel scope link src 10.10.1.99
192.168.0.0/24 dev eth1 proto kernel scope link src 192.168.0.10
172.16.0.0/24 dev eth2 proto kernel scope link src 172.16.0.55
As you can see, the default route points to 10.10.1.1, which is the gateway of the
default L3 network. Meanwhile, the /etc/resolv.conf of the VM instance is as
follows:
That is, the DNS domain is also from the default L3 network. In addition, the DNS
name server is also set to the gateway 10.10.1.1 because the default L3 network
provides a DNS server. Then, the Full Qualified Domain Name (FQDN) of the VM
instance would be as follows:
The UUID of the L3 network to which the NIC is
attached.
0.6
ip
The IP address.
0.6
mac
The MAC address.
0.6
hypervisorType
The hypervisor type.
0.6
netmask
The netmask.
0.6
gateway
The gateway.
0.6
metaData
The reserved field for internal use.
Yes
0.6
deviceId
The device ID, which is an integer that indicates
the order of NICs in the Ethernet device of the guest operating
system. For example, 0 usually means eth0, and 1 usually means eth1.
The allVolumes field is a list of volumes, including root volumes and data volumes.
To find the root volumes, you can iterate the list by checking the volume type or by
using the rootVolumeUuid field to match the volume UUID. A root volume will be
always with the VM instance until the VM instance is destroyed.
Hypervisor Type
The hypervisor type of a VM instance might be inherited from the hypervisor type of
an image or a host. This depends on how the VM instance is created.
For a VM instance that is created from RootVolumeTemplate:
Images already
have operating system installed. Therefore, the VM instance will be
created on a host with the same hypervisor type as the VM instance
image. In this case, the hypervisor type of the VM instance is inherited
from image.
For a VM instance that is created from ISO:
ISO is used to install the
operating system on a blank root volume of the VM instance. Therefore,
the VM instance might be created on a host of any hypervisor type. In
this case, the hypervisor type of the VM instance is inherited from
host.
State
A VM instance has the following states during its lifecycle:
Created
Indicates that the VM instance is just created as a record in the
database and is not started on any host. This state only exists when you
create a new VM instance.
Starting
Indicates that the VM instance is being started on a
host.
Running
Indicates that the VM instance is running on a host.
Stopping
Indicates that the VM instance is being stopped on a
host.
Stopped
Indicates that the VM instance is stopped and is not running on
any host.
Rebooting
Indicates that the VM instance is being rebooted on the host
where the VM instance was running last time.
Destroying
Indicates that the VM instance is being deleted.
Destroyed
Indicates that the VM instance is deleted
Migrating
Indicates that the VM instance is being migrated to the other
host.
Expunging
Indicates that the database of the VM instance is being
deleted.
Pausing
Indicates that the VM instance is being paused on a host.
Paused
Indicates that the VM instance is paused.
Resuming
Indicates that the VM instance is being resumed from the Paused
state to the Running state.
VolumeMigrating
Indicates that the root volume of the VM instance is being
migrated from local storage to the other host.
Unknown
Indicates that ZStack Cloud fails to detect
the state due to reasons such as losing disconnection to the
host.
The state transition diagram of a VM instance is shown in Figure 2.Figure 2. VM Instance State Transition Diagram
ZStack Cloud uses VM Tracer to periodically track the state of
VM instances. The default interval is 60 seconds. The state of a VM instance might
be affected by factors outside ZStack Cloud.
For example, the power outage of a host will cause all VM instances running
on the host to stop running. Once VM Tracer detects a mismatch between the
actual state of a VM instance and the recorded state in the database, VM
Tracer will update the database to catch up with the actual state.
If VM Tracer fails to check the state of a VM instance due to reasons such
as disconnection between the management node and the host, VM Tracer will
set the state of the VM instance to Unknown. Once VM Tracer successfully
detects the VM instance state again, for example, after the management node
is reconnected to the host, VM Tracer will update the VM instance to the
actual state.
